diff --git a/404.html b/404.html
index 448ddf0d3295..528050656254 100644
--- a/404.html
+++ b/404.html
@@ -1 +1 @@
- Page not found | Imperial NLP
Looks like there has been a mistake. Nothing exists here.
You will be redirected to the main page within 3 seconds. If not redirected, please click here.
\ No newline at end of file
+ Page not found | Imperial NLP
Looks like there has been a mistake. Nothing exists here.
You will be redirected to the main page within 3 seconds. If not redirected, please click here.
\ No newline at end of file
diff --git a/assets/bibliography/2018-12-22-distill.bib b/assets/bibliography/2018-12-22-distill.bib
deleted file mode 100644
index 2b06f3c99f2a..000000000000
--- a/assets/bibliography/2018-12-22-distill.bib
+++ /dev/null
@@ -1,7 +0,0 @@
-@article{gregor2015draw,
- title={DRAW: A recurrent neural network for image generation},
- author={Gregor, Karol and Danihelka, Ivo and Graves, Alex and Rezende, Danilo Jimenez and Wierstra, Daan},
- journal={arXiv preprint, arXiv:1502.04623},
- year={2015},
- url={https://arxiv.org/pdf/1502.04623.pdf}
-}
diff --git a/assets/downloads/NLPLists(People).csv b/assets/downloads/NLPLists(People).csv
index b708a8f3c64d..86e6bb78bb20 100644
--- a/assets/downloads/NLPLists(People).csv
+++ b/assets/downloads/NLPLists(People).csv
@@ -1,5 +1,10 @@
Name,Email,Room,Short Snippet,Longer Bio,Position,Status,Picture,Supervisor,Website,Github,Linkedin,ORCID,Google Scholar
Nikolai Rozanov,nikolai.rozanov13@imperial.ac.uk,Huxley 558,CS PhD Student in the Computing Department working on LLMs and LLM Agents.,"Nikolai is a PhD student with Dr Marek Rei. He works on LLM Agents.
Previously, Nikolai has co-founded an NLP start-up.",PhD,PhD,nikolai.jpg,Dr Marek Rei,nikolairozanov.com,ai-nikolai,nikolai-rozanov-ai,0000-0003-0274-8832,fi-feOEAAAAJ
-Dr Marek Rei,marek.rei@imperial.ac.uk,Huxley 555,Marek is a Seniour Lecturer at Imperial College London.,"Seniour Lecturer in NLP at Imperial College London. PhD from Cambridge University.
-Staff in NLP. Previously, Marek was at Cambridge.",Seniour Lecturer in NLP,Staff,marek.png,,,,,,
+Dr Marek Rei,marek.rei@imperial.ac.uk,Huxley 555,Marek is a Seniour Lecturer at Imperial College London.,"I am a researcher in Machine Learning and Natural Language Processing. My work is focused on improving machine learning architectures for representation learning, transfer learning, autoregressive modeling and multi-task optimization. Most of my research is applied in the area of Natural Language Understanding and on tasks that benefit from capturing the semantics in text, such as structured prediction, language modeling, grammatical error detection, sentiment analysis and text classification.
+
+I am a Senior Lecturer of Machine Learning at Imperial College London and a Visiting Researcher at the University of Cambridge. I am an AI Advisor for Gotya Technologies and Esgrid Technologies. I also provide consultancy services through Perception Labs.",Seniour Lecturer in NLP,Staff,marek.png,,,,,,
+Prof. Lucia Specia,l.specia@imperial.ac.uk,Huxley 572A,Lucia is a Professor in NLP at Imperial College London,"My research focuses on various aspects of data-driven approaches to natural language processing, with a particular interest in multimodal and multilingual context models and work at the intersection of language and vision. My work has various applications such as machine translation, image captioning, quality estimation and text adaptation. I currently hold an ERC (European Research Council) Starting Grant on Multi-modal Context Modelling for Machine Translation. I am also part-time Professor of Language Engineering at the University of Sheffield. Check my Sheffield page for more.",Professor in NLP,Staff,lucia.jpeg,,,,,,
+Joe Stacey,j.stacey20@imperial.ac.uk,Huxley 558,"CS PhD Student in the Computing Department working on NLP, Robustness and Explainability","I am a PhD student in Natural Language Processing, supervised by Marek Rei. My PhD is focussing on creating more robust Natural Language Inference (NLI) models that generalise better to other unseen NLI datasets. Prior to starting the PhD I worked for 6 years as a Consultant at PwC/Strategy&, and recently completed an MSc in Data Science at UCL. I also have a BSc in Mathematics from the University of Warwick, and spent two years as a Mathematics teacher on the Teach First programme.",PhD,PhD,joe.jpg,Dr Marek Rei,,,,,
+Lisa Alazraki,lisa.alazraki20@imperial.ac.uk,Huxley 558,CS PhD in Comptuing Department working on LLM Reasoning,"I am a PhD student at the Department of Computing of Imperial College London, supervised by Dr Marek Rei. I am fascinated by large foundation models and work on enhancing their planning, reasoning, and general problem-solving abilities.",PhD,PhD,lisa.png,Dr Marek Rei,,,,,
+Dr Tyler Bonnet,t.bonnet24@imperial.ac.uk,Huxley 558,CS PhD in Comptuing Department working on NLP,Tyler is pursuing his second PhD degree in the Computing Department at Imperial College London. Dr Tyler Bonnet is also Research Assistant in Natural Language Processing and Crowd Sourcing at the Oxford e-Research Centre and a PhD student in Computing at Imperial College London. He previously served as a Postdoctoral Researcher in Linked Data for Cultural Informatics at the Oxford e-Research Centre and as a Research Fellow at the Neuropolitics Research Lab of the University of Edinburgh. Tyler holds a PhD in International Relations from the London School of Economics and Political Science.,PhD,PhD,tyler.jpg,Dr Marek Rei,,,,,
diff --git a/assets/downloads/README.md b/assets/downloads/README.md
index 41a0b1831f6b..d97804688d2b 100644
--- a/assets/downloads/README.md
+++ b/assets/downloads/README.md
@@ -13,4 +13,22 @@ python3 convert.py
3. IF READY TO OVERRIDE PRODUCTION:
```bash
python3 convert.py production
+```
+
+4. Generate BIB entries:
+```bash
+python3 transform_bibtext.py production
+```
+
+
+### Testing website locally:
+
+1. Installation
+```bash
+bundle install
+```
+
+2. Running locally:
+```bash
+bundle exec jekyll serve
```
\ No newline at end of file
diff --git a/assets/downloads/convert.py b/assets/downloads/convert.py
index 8893b769117c..aa37e405ae1d 100644
--- a/assets/downloads/convert.py
+++ b/assets/downloads/convert.py
@@ -15,7 +15,7 @@ def transform_tsv_to_yaml(input_file, output_dir='_members'):
with open(input_file, 'r', encoding='utf-8') as tsv_file:
tsv_reader = csv.DictReader(tsv_file) #, delimiter='\t') #in case we want tsv file instead.
-
+
# Process each row (assuming single row in this case)
for row in tsv_reader:
# Generate the output
@@ -23,7 +23,7 @@ def transform_tsv_to_yaml(input_file, output_dir='_members'):
layout: about
inline: false
group: {row['Status']}
-group_rank: 2
+group_rank: {1 if "marek rei" in row["Name"].lower() else 2}
title: {row['Name']}
lastname: {row['Name'].split()[-1]}
diff --git a/assets/downloads/run.sh b/assets/downloads/run.sh
new file mode 100755
index 000000000000..69e5a23dd361
--- /dev/null
+++ b/assets/downloads/run.sh
@@ -0,0 +1,2 @@
+ mv ~/Downloads/NLPLists\(People\).csv .
+ python3 convert.py production
\ No newline at end of file
diff --git a/assets/downloads/transform_bibtex.py b/assets/downloads/transform_bibtex.py
new file mode 100644
index 000000000000..d6130c021840
--- /dev/null
+++ b/assets/downloads/transform_bibtex.py
@@ -0,0 +1,168 @@
+import re
+import os
+import sys
+
+def extract_arxiv_number(text):
+ """
+ Extract only the arXiv number (YYMM.NNNNN) from text
+
+ Args:
+ text (str): The text to search for an arXiv number
+
+ Returns:
+ str or None: The matched arXiv number or None if no match
+ """
+ # Regex pattern to match just the arXiv number format
+ pattern = r'\d{4}\.\d{5}'
+
+ # Search for the pattern in the text
+ match = re.search(pattern, text)
+
+ return match.group(0) if match else None
+
+def transform_bibtex_entry(original_entry):
+ """
+ Transform a single BibTeX entry by adding html, abbr, and bibtex_show fields.
+
+ Args:
+ original_entry (str): The original BibTeX entry
+
+ Returns:
+ str: Transformed BibTeX entry
+ """
+ # Remove trailing comma and newline at the end of the entry
+ original_entry = original_entry.strip().rstrip(',')
+
+ # Extract the journal for the abbr field
+ # journal_match = re.search(r'journal\s*=\s*{([^}]+)}', original_entry)
+ # abbr = journal_match.group(1).strip() if journal_match else ''
+
+ # Extract the URL for the html field
+ url_match = re.search(r'url\s*=\s*{([^}]+)}', original_entry)
+ html = url_match.group(1).strip() if url_match else ''
+ if not html:
+ arxiv_number = extract_arxiv_number(original_entry)
+ if arxiv_number:
+ html = f"https://arxiv.org/abs/{arxiv_number}"
+
+
+ # Prepare the new fields to insert
+ new_fields = [
+ ' html = {' + html + '},',
+ # f' abbr = {{{abbr}}}',
+ ' abbr = { NLP },',
+ ' bibtex_show = {true},'
+ ]
+
+ # Split the original entry into lines
+ lines = original_entry.split('\n')
+
+ # Find the index to insert new fields (after the first line)
+ insert_index = 1
+
+ # Insert the new fields
+ for new_field in reversed(new_fields):
+ lines.insert(insert_index, new_field)
+
+ # Rejoin the lines
+ transformed_entry = '\n'.join(lines)
+
+ return transformed_entry
+
+def extract_bibtex_entries(input_file):
+ """
+ Extract BibTeX entries from a single file.
+
+ Args:
+ input_file (str): Path to the input BibTeX file
+
+ Returns:
+ list: List of BibTeX entries
+ """
+ try:
+ with open(input_file, 'r', encoding='utf-8') as f:
+ file_contents = f.read()
+ except Exception as e:
+ print(f"Error reading file {input_file}: {e}")
+ return []
+
+ # Split the file into individual entries
+ # This regex looks for entries starting with @ and ending with a closing }
+ pattern = r'@(?:[^@]*})'
+ entries = re.findall(pattern, file_contents, re.DOTALL)
+
+ return entries
+
+def accumulate_and_transform_bibtex(input_directory, output_file):
+ """
+ Accumulate and transform BibTeX entries from all .bib files in a directory.
+
+ Args:
+ input_directory (str): Path to the input directory containing .bib files
+ output_file (str): Path to the output accumulated and transformed BibTeX file
+
+ Returns:
+ list: Paths to all processed files
+ """
+ # Validate input directory
+ if not os.path.isdir(input_directory):
+ print(f"Error: {input_directory} is not a valid directory.")
+ return []
+
+ # List to store all entries and processed files
+ all_entries = []
+ processed_files = []
+
+ # Iterate through all files in the input directory
+ for filename in os.listdir(input_directory):
+ # Check if the file is a .bib file
+ if filename.endswith('.bib'):
+ input_file = os.path.join(input_directory, filename)
+
+ # Extract entries from the file
+ entries = extract_bibtex_entries(input_file)
+ print(f"Len entries in file:{len(entries)}")
+
+ # Transform entries
+ transformed_entries = [transform_bibtex_entry(entry) for entry in entries]
+
+ # Add to accumulated entries
+ all_entries.extend(transformed_entries)
+ processed_files.append(input_file)
+
+ # Write accumulated entries to the output file
+ try:
+ with open(output_file, 'w', encoding='utf-8') as f:
+ f.write('\n\n'.join(all_entries))
+
+ print(f"Accumulated and transformed entries from {len(processed_files)} files.")
+ print(f"Total entries: {len(all_entries)}")
+ print(f"Output written to {output_file}")
+ except Exception as e:
+ print(f"Error writing to file {output_file}: {e}")
+ return []
+
+ return processed_files
+
+def main():
+ input_folder=os.path.join("..","bibliography")
+ output_file = os.path.join("_bibliography","papers.bib")
+
+ if len(sys.argv) == 2:
+ if sys.argv[1] == "production":
+ output_file = os.path.join("..","..",output_file)
+
+ accumulate_and_transform_bibtex(input_folder, output_file)
+
+if __name__ == '__main__':
+ main()
+# bibentry="""
+# @inproceedings{collins2018evolutionary,
+# title={Evolutionary Data Measures: Understanding the Difficulty of Text Classification Tasks},
+# author={Collins, Edward and Rozanov, Nikolai and Zhang, Bingbing},
+# booktitle={Proceedings of the 22nd Conference on Computational Natural Language Learning},
+# pages={380--391},
+# year={2018}
+# }
+# """
+# print(transform_bibtex_entry(bibentry))
\ No newline at end of file
diff --git a/assets/img/imperial-other-image-1400.webp b/assets/img/imperial-other-image-1400.webp
new file mode 100644
index 000000000000..026a88747f44
Binary files /dev/null and b/assets/img/imperial-other-image-1400.webp differ
diff --git a/assets/img/imperial-other-image-480.webp b/assets/img/imperial-other-image-480.webp
new file mode 100644
index 000000000000..211496283698
Binary files /dev/null and b/assets/img/imperial-other-image-480.webp differ
diff --git a/assets/img/imperial-other-image-800.webp b/assets/img/imperial-other-image-800.webp
new file mode 100644
index 000000000000..026a88747f44
Binary files /dev/null and b/assets/img/imperial-other-image-800.webp differ
diff --git a/assets/img/imperial-other-image.jpg b/assets/img/imperial-other-image.jpg
new file mode 100644
index 000000000000..4b57a4161103
Binary files /dev/null and b/assets/img/imperial-other-image.jpg differ
diff --git a/assets/img/team/joe-1400.webp b/assets/img/team/joe-1400.webp
new file mode 100644
index 000000000000..e167f5f5985e
Binary files /dev/null and b/assets/img/team/joe-1400.webp differ
diff --git a/assets/img/team/joe-480.webp b/assets/img/team/joe-480.webp
new file mode 100644
index 000000000000..859a540d2633
Binary files /dev/null and b/assets/img/team/joe-480.webp differ
diff --git a/assets/img/team/joe-800.webp b/assets/img/team/joe-800.webp
new file mode 100644
index 000000000000..e167f5f5985e
Binary files /dev/null and b/assets/img/team/joe-800.webp differ
diff --git a/assets/img/team/joe.jpg b/assets/img/team/joe.jpg
new file mode 100644
index 000000000000..2ba283a83345
Binary files /dev/null and b/assets/img/team/joe.jpg differ
diff --git a/assets/img/team/lisa-1400.webp b/assets/img/team/lisa-1400.webp
new file mode 100644
index 000000000000..d0ccc52053d4
Binary files /dev/null and b/assets/img/team/lisa-1400.webp differ
diff --git a/assets/img/team/lisa-480.webp b/assets/img/team/lisa-480.webp
new file mode 100644
index 000000000000..3cad49e2cb54
Binary files /dev/null and b/assets/img/team/lisa-480.webp differ
diff --git a/assets/img/team/lisa-800.webp b/assets/img/team/lisa-800.webp
new file mode 100644
index 000000000000..d0ccc52053d4
Binary files /dev/null and b/assets/img/team/lisa-800.webp differ
diff --git a/assets/img/team/lisa.png b/assets/img/team/lisa.png
new file mode 100644
index 000000000000..ab82b4ac22d1
Binary files /dev/null and b/assets/img/team/lisa.png differ
diff --git a/assets/img/team/lucia-1400.webp b/assets/img/team/lucia-1400.webp
new file mode 100644
index 000000000000..812b97e8bc5f
Binary files /dev/null and b/assets/img/team/lucia-1400.webp differ
diff --git a/assets/img/team/lucia-480.webp b/assets/img/team/lucia-480.webp
new file mode 100644
index 000000000000..9cf39f410ca2
Binary files /dev/null and b/assets/img/team/lucia-480.webp differ
diff --git a/assets/img/team/lucia-800.webp b/assets/img/team/lucia-800.webp
new file mode 100644
index 000000000000..812b97e8bc5f
Binary files /dev/null and b/assets/img/team/lucia-800.webp differ
diff --git a/assets/img/team/lucia.jpeg b/assets/img/team/lucia.jpeg
new file mode 100644
index 000000000000..a0a234160c28
Binary files /dev/null and b/assets/img/team/lucia.jpeg differ
diff --git a/assets/img/team/nikolai-1400.webp b/assets/img/team/nikolai-1400.webp
index 40569e6db19c..69a8726ec7dd 100644
Binary files a/assets/img/team/nikolai-1400.webp and b/assets/img/team/nikolai-1400.webp differ
diff --git a/assets/img/team/nikolai-480.webp b/assets/img/team/nikolai-480.webp
index 9bd6916952e7..fd1491faa570 100644
Binary files a/assets/img/team/nikolai-480.webp and b/assets/img/team/nikolai-480.webp differ
diff --git a/assets/img/team/nikolai-800.webp b/assets/img/team/nikolai-800.webp
index 40569e6db19c..69a8726ec7dd 100644
Binary files a/assets/img/team/nikolai-800.webp and b/assets/img/team/nikolai-800.webp differ
diff --git a/assets/img/team/nikolai.jpg b/assets/img/team/nikolai.jpg
index 2b8943822e89..39273cc18302 100644
Binary files a/assets/img/team/nikolai.jpg and b/assets/img/team/nikolai.jpg differ
diff --git a/assets/img/team/tyler-1400.webp b/assets/img/team/tyler-1400.webp
new file mode 100644
index 000000000000..edec1a3790d3
Binary files /dev/null and b/assets/img/team/tyler-1400.webp differ
diff --git a/assets/img/team/tyler-480.webp b/assets/img/team/tyler-480.webp
new file mode 100644
index 000000000000..faa291d6ee0a
Binary files /dev/null and b/assets/img/team/tyler-480.webp differ
diff --git a/assets/img/team/tyler-800.webp b/assets/img/team/tyler-800.webp
new file mode 100644
index 000000000000..edec1a3790d3
Binary files /dev/null and b/assets/img/team/tyler-800.webp differ
diff --git a/assets/img/team/tyler.jpg b/assets/img/team/tyler.jpg
new file mode 100644
index 000000000000..b8085fe4028d
Binary files /dev/null and b/assets/img/team/tyler.jpg differ
diff --git a/assets/stash/cope collab.html b/assets/stash/cope collab.html
index 07b0d2a3fce8..33802854444f 100644
--- a/assets/stash/cope collab.html
+++ b/assets/stash/cope collab.html
@@ -1 +1 @@
- | Imperial NLP
\ No newline at end of file
+ | Imperial NLP
\ No newline at end of file
diff --git a/assets/stash/cope member.html b/assets/stash/cope member.html
index 1ed9c1edb85e..0a446b3bcffa 100644
--- a/assets/stash/cope member.html
+++ b/assets/stash/cope member.html
@@ -1 +1 @@
- Leslie S. Cope, PhD | Imperial NLP
Leslie M. Cope, PhD Associate Professor of Oncology - Biostatistics and Bioinformatics
Leslie Cope, Ph.D., is a biostatistician and bioinformatician with expertise in the design and analysis of high-throughput genomic data.
He co-directs the Johns Hopkins Kimmel Cancer Center’s Experimental and Computational Genomics Core. Cope also has participated in research teams including Stand Up to Cancer and the Department of Defense’s Center of Excellence in Ovarian Cancer.
Cope is a member of the Epigenetics Genome Characterization program within the National Institutes of Health-funded project, The Cancer Genome Atlas (TCGA), where he leads DNA methylation data analysis and its integration with other data types for several TCGA disease working groups. His research focuses on methods for identifying and validating biomarkers, including tools for integrating multiple sources of data to take advantage of the vast stores of archival data that are now available.
Short Bio
Dr. Cope received his master’s degree and Ph.D. from The Johns Hopkins University and is a member of the American Statistical Association, International Biometric Society and Institute of Mathematical Statistics.
\ No newline at end of file
+ Leslie S. Cope, PhD | Imperial NLP
Leslie M. Cope, PhD Associate Professor of Oncology - Biostatistics and Bioinformatics
Leslie Cope, Ph.D., is a biostatistician and bioinformatician with expertise in the design and analysis of high-throughput genomic data.
He co-directs the Johns Hopkins Kimmel Cancer Center’s Experimental and Computational Genomics Core. Cope also has participated in research teams including Stand Up to Cancer and the Department of Defense’s Center of Excellence in Ovarian Cancer.
Cope is a member of the Epigenetics Genome Characterization program within the National Institutes of Health-funded project, The Cancer Genome Atlas (TCGA), where he leads DNA methylation data analysis and its integration with other data types for several TCGA disease working groups. His research focuses on methods for identifying and validating biomarkers, including tools for integrating multiple sources of data to take advantage of the vast stores of archival data that are now available.
Short Bio
Dr. Cope received his master’s degree and Ph.D. from The Johns Hopkins University and is a member of the American Statistical Association, International Biometric Society and Institute of Mathematical Statistics.
\ No newline at end of file
diff --git a/assets/stash/easwaran collab.html b/assets/stash/easwaran collab.html
index 323e643a709c..23ea4f370fcb 100644
--- a/assets/stash/easwaran collab.html
+++ b/assets/stash/easwaran collab.html
@@ -1 +1 @@
- | Imperial NLP
\ No newline at end of file
+ | Imperial NLP
\ No newline at end of file
diff --git a/assets/stash/easwaran member.html b/assets/stash/easwaran member.html
index 78e6f4668afd..6b081716d0ed 100644
--- a/assets/stash/easwaran member.html
+++ b/assets/stash/easwaran member.html
@@ -1 +1 @@
- Hariharan 'Hari' Easwaran, PhD, MSc | Imperial NLP
Hariharan P. Easwaran, PhD, MSc Associate Professor of Oncology - Cancer Biology
\ No newline at end of file
diff --git a/feed.xml b/feed.xml
index fe16cc4d4e35..65a973b4edf8 100644
--- a/feed.xml
+++ b/feed.xml
@@ -1 +1 @@
-Jekyll2024-12-12T03:32:37+00:00https://hunky-d0ry.github.io/feed.xmlImperial NLPImperial NLP website. Displaying External Posts on Your al-folio Blog2022-04-23T23:20:09+00:002022-04-23T23:20:09+00:00https://hunky-d0ry.github.io/2022/04/23/displaying-external-posts-on-your-al-folio-blog
\ No newline at end of file
+Jekyll2024-12-13T03:27:28+00:00https://hunky-d0ry.github.io/feed.xmlImperial NLPImperial NLP website. Displaying External Posts on Your al-folio Blog2022-04-23T23:20:09+00:002022-04-23T23:20:09+00:00https://hunky-d0ry.github.io/2022/04/23/displaying-external-posts-on-your-al-folio-blog
\ No newline at end of file
diff --git a/index.html b/index.html
index 2f70018f318f..1d72145ede4e 100644
--- a/index.html
+++ b/index.html
@@ -1 +1 @@
- Imperial NLP
Welcome to the NLP Group website.
\ No newline at end of file
+ Imperial NLP
NLP Group @ Imperial College London
The NLP Group aims to develop machine learning algorithms for language understanding and generation following a holistic approach where language can be combined with other types of information. In addition to textual context of varying length, our models also sometimes include knowledge in the form of images, videos, audio recordings, sensor data, metadata and knowledge bases.
Machine learning is at the core of our research. The main areas we work on include large language models (LLMs), robustness & generalisation, LLM reasoning, LLM agents, language generation, machine translation, text adaptation, transfer learning for language, NLP for health and sustainable development, quality evaluation and estimation, and language learning applications.
Topics we work on in NLP 🌍💬
LLLM Reasoning & Agents 🤖🔬
Natural Language Inference 🧠🔍
Robustness & Generalisation 💪🌐
Explainability in NLP 💡🗣️
NLP for Healthcare ❤️👩⚕️
\ No newline at end of file
diff --git a/members/joe.html b/members/joe.html
new file mode 100644
index 000000000000..c7a6d9dd6afc
--- /dev/null
+++ b/members/joe.html
@@ -0,0 +1 @@
+ Joe Stacey | Imperial NLP
I am a PhD student in Natural Language Processing, supervised by Marek Rei. My PhD is focussing on creating more robust Natural Language Inference (NLI) models that generalise better to other unseen NLI datasets. Prior to starting the PhD I worked for 6 years as a Consultant at PwC/Strategy&, and recently completed an MSc in Data Science at UCL. I also have a BSc in Mathematics from the University of Warwick, and spent two years as a Mathematics teacher on the Teach First programme.
\ No newline at end of file
diff --git a/members/lisa.html b/members/lisa.html
new file mode 100644
index 000000000000..50b793060c88
--- /dev/null
+++ b/members/lisa.html
@@ -0,0 +1 @@
+ Lisa Alazraki | Imperial NLP
I am a PhD student at the Department of Computing of Imperial College London, supervised by Dr Marek Rei. I am fascinated by large foundation models and work on enhancing their planning, reasoning, and general problem-solving abilities.
\ No newline at end of file
diff --git a/members/marek.html b/members/marek.html
index 7bd9dac8675f..55aa228f970f 100644
--- a/members/marek.html
+++ b/members/marek.html
@@ -1 +1 @@
- Dr Marek Rei | Imperial NLP
I am a researcher in Machine Learning and Natural Language Processing. My work is focused on improving machine learning architectures for representation learning, transfer learning, autoregressive modeling and multi-task optimization. Most of my research is applied in the area of Natural Language Understanding and on tasks that benefit from capturing the semantics in text, such as structured prediction, language modeling, grammatical error detection, sentiment analysis and text classification.
I am a Senior Lecturer of Machine Learning at Imperial College London and a Visiting Researcher at the University of Cambridge. I am an AI Advisor for Gotya Technologies and Esgrid Technologies. I also provide consultancy services through Perception Labs.
\ No newline at end of file
diff --git a/members/nikolai.html b/members/nikolai.html
index dfd42f2a4fd6..6f6aca0a7f2f 100644
--- a/members/nikolai.html
+++ b/members/nikolai.html
@@ -1 +1 @@
- Nikolai Rozanov | Imperial NLP
Nikolai is a PhD student with Dr Marek Rei. He works on LLM Agents. Previously, Nikolai has co-founded an NLP start-up.
\ No newline at end of file
diff --git a/members/prof.html b/members/prof.html
new file mode 100644
index 000000000000..327f724092fb
--- /dev/null
+++ b/members/prof.html
@@ -0,0 +1 @@
+ Prof. Lucia Specia | Imperial NLP
My research focuses on various aspects of data-driven approaches to natural language processing, with a particular interest in multimodal and multilingual context models and work at the intersection of language and vision. My work has various applications such as machine translation, image captioning, quality estimation and text adaptation. I currently hold an ERC (European Research Council) Starting Grant on Multi-modal Context Modelling for Machine Translation. I am also part-time Professor of Language Engineering at the University of Sheffield. Check my Sheffield page for more.
\ No newline at end of file
diff --git a/members/tyler.html b/members/tyler.html
new file mode 100644
index 000000000000..b95cafae1ca4
--- /dev/null
+++ b/members/tyler.html
@@ -0,0 +1 @@
+ Dr Tyler Bonnet | Imperial NLP
Tyler is pursuing his second PhD degree in the Computing Department at Imperial College London. Dr Tyler Bonnet is also Research Assistant in Natural Language Processing and Crowd Sourcing at the Oxford e-Research Centre and a PhD student in Computing at Imperial College London. He previously served as a Postdoctoral Researcher in Linked Data for Cultural Informatics at the Oxford e-Research Centre and as a Research Fellow at the Neuropolitics Research Lab of the University of Edinburgh. Tyler holds a PhD in International Relations from the London School of Economics and Political Science.
\ No newline at end of file
diff --git a/publications/index.html b/publications/index.html
index 5505b4906b4d..e057eca918cd 100644
--- a/publications/index.html
+++ b/publications/index.html
@@ -1,268 +1,49 @@
- publications | Imperial NLP
Publications
Publications from the NLP Group.
2023
Advanced Science
Multiplex Digital Methylation-Specific PCR for Noninvasive Screening of Lung Cancer
Zhao, Yang, O’Keefe, Christine M., Hsieh, Kuangwen, Cope, Leslie, Joyce, Sonali C., Pisanic, Thomas R., Herman, James G., and Wang, Tza-Huei
Abstract There remains tremendous interest in developing liquid biopsy assays for detection of cancer-specific alterations, such as mutations and DNA methylation, in cell-free DNA (cfDNA) obtained through noninvasive blood draws. However, liquid biopsy analysis is often challenging due to exceedingly low fractions of circulating tumor DNA (ctDNA), necessitating the use of extended tumor biomarker panels. While multiplexed PCR strategies provide advantages such as higher throughput, their implementation is often hindered by challenges such as primer-dimers and PCR competition. Alternatively, digital PCR (dPCR) approaches generally offer superior performance, but with constrained multiplexing capability. This paper describes development and validation of the first multiplex digital methylation-specific PCR (mdMSP) platform for simultaneous analysis of four methylation biomarkers for liquid-biopsy-based detection of non-small cell lung cancer (NSCLC). mdMSP employs a microfluidic device containing four independent, but identical modules, housing a total of 40 160 nanowells. Analytical validation of the mdMSP platform demonstrates multiplex detection at analytical specificities as low as 0.0005%. The clinical utility of mdMSP is also demonstrated in a cohort of 72 clinical samples of low-volume liquid biopsy specimens from patients with computed tomography (CT)-scan indeterminant pulmonary nodules, exhibiting superior clinical performance when compared to traditional MSP assays for noninvasive detection of early-stage NSCLC.
@article{RN80,
- author={Zhao, Yang and O'Keefe, Christine M. and Hsieh, Kuangwen and Cope, Leslie and Joyce, Sonali C. and Pisanic, Thomas R. and Herman, James G. and Wang, Tza-Huei},
- title={Multiplex Digital Methylation-Specific PCR for Noninvasive Screening of Lung Cancer},
- journal={Advanced Science},
- volume={n/a},
- number={n/a},
- pages={2206518},
- issn={2198-3844},
- doi={https://doi.org/10.1002/advs.202206518},
- url={https://onlinelibrary.wiley.com/doi/abs/10.1002/advs.202206518},
+ publications | Imperial NLP
Publications
Publications from the NLP Group.
2024
NLP
Efficient Exploration in Deep Reinforcement Learning: A Novel Bayesian Actor-Critic Algorithm
@article{rozanov2024stateact,
+ title={StateAct: State Tracking and Reasoning for Acting and Planning with Large Language Models},
+ author={Rozanov, Nikolai and Rei, Marek},
+ journal={arXiv preprint arXiv:2410.02810},
+ year={2024}
+}
NLP
IsoChronoMeter: A simple and effective isochronic translation evaluation metric
Rozanov, Nikolai, Pankov, Vikentiy, Mukhutdinov, Dmitrii, and Vypirailenko, Dima
@article{rozanov2024isochronometer,
+ title={IsoChronoMeter: A simple and effective isochronic translation evaluation metric},
+ author={Rozanov, Nikolai and Pankov, Vikentiy and Mukhutdinov, Dmitrii and Vypirailenko, Dima},
+ journal={arXiv preprint arXiv:2410.11127},
+ year={2024}
+}
2023
NLP
Learning From Free-Text Human Feedback–Collect New Datasets Or Extend Existing Ones?
Petrak, Dominic, Moosavi, Nafise Sadat, Tian, Ye, Rozanov, Nikolai, and Gurevych, Iryna
@article{petrak2023learning,
+ title={Learning From Free-Text Human Feedback--Collect New Datasets Or Extend Existing Ones?},
+ author={Petrak, Dominic and Moosavi, Nafise Sadat and Tian, Ye and Rozanov, Nikolai and Gurevych, Iryna},
+ journal={arXiv preprint arXiv:2310.15758},year={2023}
-}
2022
MicroTAS 2022
A microfluidic platform for quantitative multiplex profiling of DNA methylation biomarkers.
Zhao, Y, O’Keefe, C. M., Herman, J., Pisanic, T.R., and Wang, T. H.
The 26th International Conference on Miniaturized Systems for Chemistry and Life Sciences (MicroTAS 2022), 2022
@article{RN79,
- author={Zhao, Y and O’Keefe, C. M. and Herman, J. and Pisanic, T.R. and Wang, T. H.},
- title={A microfluidic platform for quantitative multiplex profiling of DNA methylation biomarkers. },
- journal={The 26th International Conference on Miniaturized Systems for Chemistry and Life Sciences (MicroTAS 2022)},
+}
2022
NLP
Connecting the Semantic Dots: Zero-Shot Learning with Self-Aligning Autoencoders and a New Contrastive-Loss for Negative Sampling
Terry-Jack, Mohammed, and Rozanov, Nikolai
In 21st IEEE International Conference on Machine Learning and Applications, 2022
@inproceedings{terry2022connecting,
+ author={Terry-Jack, Mohammed and Rozanov, Nikolai},
+ booktitle={21st IEEE International Conference on Machine Learning and Applications},
+ pages={1504--1511},year={2022}
-}
Cancers
Current and Emerging Methods for Ovarian Cancer Screening and Diagnostics: A Comprehensive Review
Liberto, Juliane M., Chen, Sheng-Yin, Shih, Ie-Ming, Wang, Tza-Huei, Wang, Tian-Li, and Pisanic, Thomas R.
Ovarian high-grade serous carcinoma (HGSC) has a 5-year survival rate of less than 50%, making it one of the most lethal gynecological cancers for women in the developed world today. Delayed presentation of clinical symptoms and late-stage diagnosis drive the high mortality rate of this disease. Early detection is associated with significant improvements in survival, however, screening in the general population is currently not recommended at this time due to a notable lack of sensitive and specific biomarkers for early-stage disease. In this review, we provide an overview of the current landscape of ovarian cancer diagnostics, emphasizing emerging methodologies for the non-invasive detection of HGSC.
@article{RN74,
- author={Liberto, Juliane M. and Chen, Sheng-Yin and Shih, Ie-Ming and Wang, Tza-Huei and Wang, Tian-Li and Pisanic, Thomas R.},
- title={Current and Emerging Methods for Ovarian Cancer Screening and Diagnostics: A Comprehensive Review},
- journal={Cancers},
- volume={14},
- number={12},
- pages={2885},
- issn={2072-6694},
- url={https://www.mdpi.com/2072-6694/14/12/2885},
- year={2022},
-}
Biosens. Bioelect.
Magnetofluidic immuno-PCR for point-of-care COVID-19 serological testing
Zhang, Pengfei, Chen, Liben, Hu, Jiumei, Trick, Alexander Y., Chen, Fan-En, Hsieh, Kuangwen, Zhao, Yang, Coleman, Branch, Kruczynski, Kate, Pisanic, Thomas R., Heaney, Christopher D., Clarke, William A., and Wang, Tza-Huei
Serological tests play an important role in the fight against Coronavirus Disease 2019 (COVID-19), including monitoring the dynamic immune response after vaccination, identifying past infection and determining community infection rate. Conventional methods for serological testing, such as enzyme-linked immunosorbent assays and chemiluminescence immunoassays, provide reliable and sensitive antibody detection but require sophisticated laboratory infrastructure and/or lengthy assay time. Conversely, lateral flow immunoassays are suitable for rapid point-of-care tests but have limited sensitivity. Here, we describe the development of a rapid and sensitive magnetofluidic immuno-PCR platform that can address the current gap in point-of-care serological testing for COVID-19. Our magnetofluidic immuno-PCR platform automates a magnetic bead-based, single-binding, and one-wash immuno-PCR assay in a palm-sized magnetofluidic device and delivers results in ∼30 min. In the device, a programmable magnetic arm attracts and transports magnetically-captured antibodies through assay reagents pre-loaded in a companion plastic cartridge, and a miniaturized thermocycler and a fluorescence detector perform immuno-PCR to detect the antibodies. We evaluated our magnetofluidic immuno-PCR with 108 clinical serum/plasma samples and achieved 93.8% (45/48) sensitivity and 98.3% (59/60) specificity, demonstrating its potential as a rapid and sensitive point-of-care serological test for COVID-19.
@article{RN68,
- author={Zhang, Pengfei and Chen, Liben and Hu, Jiumei and Trick, Alexander Y. and Chen, Fan-En and Hsieh, Kuangwen and Zhao, Yang and Coleman, Branch and Kruczynski, Kate and Pisanic, Thomas R. and Heaney, Christopher D. and Clarke, William A. and Wang, Tza-Huei},
- title={Magnetofluidic immuno-PCR for point-of-care COVID-19 serological testing},
- journal={Biosensors and Bioelectronics},
- volume={195},
- pages={113656},
- issn={0956-5663},
- doi={https://doi.org/10.1016/j.bios.2021.113656},
- url={https://www.sciencedirect.com/science/article/pii/S095656632100693X},
- year={2022}
-}
2021
SLAS Techn.
High-throughput sample processing for methylation analysis in an automated, enclosed environment
Stark, Alejandro, Pisanic, Thomas R., Herman, James G., and Wang, Tza-Huei
Variation in methylcytosine is perhaps the most well-studied epigenetic mechanism of gene regulation. Methods that have been developed and implemented for assessing DNA methylation require sample DNA to be extracted, purified and chemically-processed through bisulfite conversion before downstream analysis. While some automated solutions exist for each of these individual process steps, a fully integrated solution for accomplishing the entire process in a high-throughput manner has yet to be demonstrated. Thus, sample processing methods still require numerous manual steps that may reduce sample throughput and precision, while increasing the risk of contamination and human error. In this work, we present an integrated, automated solution for performing the entire sample preparation process, including DNA extraction, purification, bisulfite conversion and PCR plate preparation within in an enclosed environment. The method employs silica-coated magnetic particles that eliminate the need for a centrifuge or vacuum manifold, thereby reducing the complexity and cost of the required automation platform. Toward this end, we also compare commercial DNA extraction and bisulfite conversion kits to identify a protocol suitable for automation to significantly improve genomic and bisulfite-treated DNA yields over manufacturer protocols. Overall, this research demonstrated development of an automated protocol that offers the ability to generate high-quality, bisulfite-treated DNA samples in a high-throughput and clean environment with minimal user intervention and comparable yields to manual processing.
@article{RN72,
- author={Stark, Alejandro and Pisanic, Thomas R. and Herman, James G. and Wang, Tza-Huei},
- title={High-throughput sample processing for methylation analysis in an automated, enclosed environment},
- journal={SLAS Technology},
- issn={2472-6303},
- doi={https://doi.org/10.1016/j.slast.2021.12.002},
- url={https://www.sciencedirect.com/science/article/pii/S2472630321000248},
- year={2021}
-}
Anal. Chem.
Ligation-Enabled Fluorescence-Coding PCR for High-Dimensional Fluorescence-Based Nucleic Acid Detection
Park, Joon Soo, Pisanic, Thomas, Zhang, Ye, and Wang, Tza-Huei
Polymerase chain reaction (PCR) is by far the most commonly used method of nucleic acid amplification and has likewise been employed for a plethora of diagnostic purposes. Nonetheless, multiplexed PCR-based detection schemes have hitherto been largely limited by technical challenges associated with nonspecific interactions and other limitations inherent to traditional fluorescence-based assays. Here, we describe a novel strategy for multiplexed PCR-based analysis called Ligation-eNabled fluorescence-Coding PCR (LiNC PCR) that exponentially enhances the multiplexing capability of standard fluorescence-based PCR assays. The technique relies upon a simple, preliminary ligation reaction in which target DNA sequences are converted to PCR template molecules with distinct endpoint fluorescence signatures. Universal TaqMan probes are used to create target-specific multicolor fluorescence signals that can be readily decoded to identify amplified targets of interest. We demonstrate the LiNC PCR technique by implementing a two-color-based assay for detection of 10 ovarian cancer epigenetic biomarkers at analytical sensitivities as low as 60 template molecules with no detectable target cross-talk. Overall, LiNC PCR provides a simple and inexpensive method for achieving high-dimensional multiplexing that can be implemented in manifold molecular diagnostic applications.
@article{RN54,
- author={Park, Joon Soo and Pisanic, Thomas and Zhang, Ye and Wang, Tza-Huei},
- title={Ligation-Enabled Fluorescence-Coding PCR for High-Dimensional Fluorescence-Based Nucleic Acid Detection},
- journal={Analytical Chemistry},
- issn={0003-2700},
- doi={10.1021/acs.analchem.0c04221},
- url={https://doi.org/10.1021/acs.analchem.0c04221},
- year={2021}
-}
J. Pathol.
Mutation and methylation profiles of ectopic and eutopic endometrial tissues
@article{RN67,
- author={Li, Lihong and Antero, Maria Facadio and Zhang, Ming and Chu, Tiffany and Seckin, Tamer and Ayhan, Ayse and Pisanic, Thomas and Wang, Tian-Li and Cope, Leslie and Segars, James and Shih, Ie-Ming},
- title={Mutation and methylation profiles of ectopic and eutopic endometrial tissues},
- journal={The Journal of Pathology},
- volume={255},
- number={4},
- pages={387-398},
- issn={0022-3417},
- doi={https://doi.org/10.1002/path.5778},
- url={https://onlinelibrary.wiley.com/doi/abs/10.1002/path.5778},
- year={2021}
-}
Clin. Epigenetics
High performance methylated DNA markers for detection of colon adenocarcinoma
Klein Kranenbarg, Romy A. M., Vali, Abdul Hussain, Ijzermans, Jan N. M., Pisanic, Thomas R., Wang, Tza-Huei, Azad, Nilofer, Sukumar, Saraswati, and Fackler, Mary Jo
Colon cancer (CC) is treatable if detected in its early stages. Improved CC detection assays that are highly sensitive, specific, and available at point of care are needed. In this study, we systematically selected and tested methylated markers that demonstrate high sensitivity and specificity for detection of CC in tissue and circulating cell-free DNA.
@article{RN69,
- author={Klein Kranenbarg, Romy A. M. and Vali, Abdul Hussain and Ijzermans, Jan N. M. and Pisanic, Thomas R. and Wang, Tza-Huei and Azad, Nilofer and Sukumar, Saraswati and Fackler, Mary Jo},
- title={High performance methylated DNA markers for detection of colon adenocarcinoma},
- journal={Clinical Epigenetics},
- volume={13},
- number={1},
- pages={218},
- issn={1868-7083},
- doi={10.1186/s13148-021-01206-2},
- url={https://doi.org/10.1186/s13148-021-01206-2},
- year={2021}
-}
Clin. Transl. Sci.
A phase 2 trial of gemcitabine and docetaxel in patients with metastatic colorectal adenocarcinoma with methylated checkpoint with forkhead and ring finger domain promoter and/or microsatellite instability phenotype
Baretti, Marina, Karunasena, Enusha, Zahurak, Marianna, Walker, Rosalind, Zhao, Yang, Pisanic, Thomas R., Wang, Tza-Huei, Greten, Tim F., Duffy, Austin G., Gootjes, Elske, Meijer, Gerrit, Verheul, Henk M.W., Ahuja, Nita, Herman, James G., and Azad, Nilofer S.
Abstract We previously reported CHFR methylation in a subset of colorectal cancer (CRC; ∼30%) with high concordance with microsatellite instability (MSI). We also showed that CHFR methylation predicted for sensitivity to docetaxel, whereas the MSI-high phenotypes were sensitive to gemcitabine. We hypothesized that this subset of patients with CRC would be selectively sensitive to gemcitabine and docetaxel. We enrolled a Phase 2 trial of gemcitabine and docetaxel in patients with MSI-high and/or CHFR methylated CRC. The primary objective was Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 response rate. Enrolled patients were treated with gemcitabine 800 mg/m2 on days 1 and 8 and docetaxel 70 mg/m2 on day 8 of each 21-day cycle. A total of 6 patients with CHFR-methylated, MSI-high CRC were enrolled from September 2012 to August 2016. The study was closed in September of 2017 due to poor accrual prior to reaching the first interim assessment of response rate, which would have occurred at 10 patients. No RECIST criteria tumor responses were observed, with 3 patients (50%) having stable disease as best response, 1 lasting more than 9 months. Median progression-free survival (PFS) was 1.79 months (95% confidence interval [CI] = 1.28, not available [NA]) and median overall survival (OS) was 15.67 months (95% CI = 4.24, NA). Common grade 3 toxicities were lymphopenia (67%), leukopenia (33%), and anemia (33%). Although negative, this study establishes a proof-of-concept for the implementation of epigenetic biomarkers (CHFR methylation/MSI) as inclusion criteria in a prospective clinical trial to optimize combinatorial strategies in the era of personalized medicine.
@article{RN55,
- author={Baretti, Marina and Karunasena, Enusha and Zahurak, Marianna and Walker, Rosalind and Zhao, Yang and Pisanic, Thomas R. and Wang, Tza-Huei and Greten, Tim F. and Duffy, Austin G. and Gootjes, Elske and Meijer, Gerrit and Verheul, Henk M.W. and Ahuja, Nita and Herman, James G. and Azad, Nilofer S.},
- title={A phase 2 trial of gemcitabine and docetaxel in patients with metastatic colorectal adenocarcinoma with methylated checkpoint with forkhead and ring finger domain promoter and/or microsatellite instability phenotype},
- journal={Clinical and Translational Science},
- volume={n/a},
- number={n/a},
- issn={1752-8054},
- doi={https://doi.org/10.1111/cts.12960},
- url={https://ascpt.onlinelibrary.wiley.com/doi/abs/10.1111/cts.12960},
+}
@inproceedings{cucurnia2021matilda,
+ title={MATILDA-Multi-AnnoTator multi-language InteractiveLight-weight Dialogue Annotator},
+ author={Cucurnia, Davide and Rozanov, Nikolai and Sucameli, Irene and Ciuffoletti, Augusto and Simi, Maria},
+ booktitle={Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations},
+ pages={32--39},year={2021}
-}
2020
Clin. Epigenetics
Leveraging locus-specific epigenetic heterogeneity to improve the performance of blood-based DNA methylation biomarkers
Miller, Brendan F., Pisanic, Thomas R., Margolin, Gennady, Petrykowska, Hanna M., Athamanolap, Pornpat, Osei-Tutu, Akosua, Annunziata, Christina M., Wang, Tza-Huei, and Elnitski, Laura
Aberrant DNA methylation is commonly heralded as a promising cancer biomarker; however, its inherently stochastic nature often leads to variable methylation patterns that can complicate the use of methylation biomarkers for clinical diagnostics, particularly in dilute samples such as liquid biopsies. Here, we present a methylation density binary classifier, a statistical method for leveraging differential heterogeneous methylation to predict and optimize the performance of methylation biomarkers for clinical applications. We first developed and tested the classifier using methylation density profiles derived from reduced representation bisulfite sequencing reads of ovarian carcinoma at ZNF154, a recurrently methylated locus in multiple cancer types. We then used in silico simulations to predict the performance of the classifier in liquid biopsies and validated these predictions using quasi-digital melt curve analysis (DREAMing) of circulating cell-free DNA from individuals with versus without ovarian carcinoma. We found good agreement between predicted and observed classifier performance, and further demonstrated that implementation of this approach with ZNF154 outperformed CA-125 for use in etiologically-diverse ovarian cancer types. Our results indicate that methylation density profiles can be exploited to predict and facilitate implementation of methylation biomarkers for clinical applications, and that ZNF154 methylation shows promise as a clinically-useful biomarker for ovarian cancer.
@article{RN44,
- author={Miller, Brendan F. and Pisanic, Thomas R. and Margolin, Gennady and Petrykowska, Hanna M. and Athamanolap, Pornpat and Osei-Tutu, Akosua and Annunziata, Christina M. and Wang, Tza-Huei and Elnitski, Laura},
- title={Leveraging locus-specific epigenetic heterogeneity to improve the performance of blood-based DNA methylation biomarkers},
- journal={Clinical Epigenetics},
- volume={12},
- doi={10.1186/s13148-020-00939-w},
- url={https://www.biorxiv.org/content/10.1101/579839v2},
- year={2020},
-}
Clin. Cancer Res.
Methylomic Landscapes of Ovarian Cancer Precursor Lesions
Pisanic, Thomas R., Wang, Yeh, Sun, Hanru, Considine, Michael, Li, Lihong, Wang, Tza-Huei, Wang, Tian-Li, and Shih, Ie-Ming
Purpose: The current paradigm in the development of high-grade serous ovarian carcinoma (HGSC) proposes that the majority of HGSCs arise from precursor serous tubal intraepithelial carcinoma (STIC) lesions of the fallopian tube. Here we survey genome-wide methylation in HGSC precursor lesions to identify genomic regions that exhibit high-specificity differential hypermethylation for potential use as biomarkers for detecting STIC and HGSC at stages when curative intervention likely remains feasible.Experimental Design: We first identified quality control criteria for performing reliable methylomic analysis of DNA-limited tubal precursor lesions with the Illumina Infinium MethylationEPIC array. We then used this platform to compare genome-wide methylation among 12 STICs with paired adjacent-normal epithelia, one p53 signature lesion and two samples of concurrent HGSC. The resulting methylomic data were analyzed by unsupervised hierarchical clustering and multidimensional analysis. Regions of high-confidence STIC-specific differential hypermethylation were identified using selective bioinformatic criteria and compared with published MethylationEPIC data from 23 HGSC tumors and 11 healthy fallopian tube mucosae.Results: Unsupervised analysis showed that STICs largely clustered with HGSCs, but were clearly distinct from adjacent-normal fallopian tube epithelia. Forty-two genomic regions exhibited high-confidence STIC-specific differential hypermethylation, of which 17 (40.5%) directly overlapped with HGSC-specific differentially methylated regions. Methylation at these shared loci was able to completely distinguish STIC and HGSC samples from normal and adjacent-normal specimens.Conclusions: Our results suggest that most STICs are epigenetically similar to HGSCs and share regions of differential hypermethylation that warrant further evaluation for potential use as biomarkers for early detection of ovarian HGSC.See related commentary by Ishak and De Carvalho, p. 6083
@article{RN50,
- author={Pisanic, Thomas R. and Wang, Yeh and Sun, Hanru and Considine, Michael and Li, Lihong and Wang, Tza-Huei and Wang, Tian-Li and Shih, Ie-Ming},
- title={Methylomic Landscapes of Ovarian Cancer Precursor Lesions},
- journal={Clinical Cancer Research},
- volume={26},
- number={23},
- pages={6310-6320},
- doi={10.1158/1078-0432.Ccr-20-0270},
- year={2020},
-}
Fan, Huihui, Atiya, Huda I., Wang, Yeh, Pisanic, Thomas R., Wang, Tza-Huei, Shih, Ie-Ming, Foy, Kelly K., Frisbie, Leonard, Buckanovich, Ronald J., Chomiak, Alison A., Tiedemann, Rochelle L., Rothbart, Scott B., Chandler, Chelsea, Shen, Hui, and Coffman, Lan G.
A role for cancer cell epithelial-to-mesenchymal transition (EMT) in cancer is well established. Here, we show that, in addition to cancer cell EMT, ovarian cancer cell metastasis relies on an epigenomic mesenchymal-to-epithelial transition (MET) in host mesenchymal stem cells (MSCs). These reprogrammed MSCs, termed carcinoma-associated MSCs (CA-MSCs), acquire pro-tumorigenic functions and directly bind cancer cells to serve as a metastatic driver/chaperone. Cancer cells induce this epigenomic MET characterized by enhancer-enriched DNA hypermethylation, altered chromatin accessibility, and differential histone modifications. This phenomenon appears clinically relevant, as CA-MSC MET is highly correlated with patient survival. Mechanistically, mirroring MET observed in development, MET in CA-MSCs is mediated by WT1 and EZH2. Importantly, EZH2 inhibitors, which are clinically available, significantly inhibited CA-MSC-mediated metastasis in mouse models of ovarian cancer.
@article{RN53,
- author={Fan, Huihui and Atiya, Huda I. and Wang, Yeh and Pisanic, Thomas R. and Wang, Tza-Huei and Shih, Ie-Ming and Foy, Kelly K. and Frisbie, Leonard and Buckanovich, Ronald J. and Chomiak, Alison A. and Tiedemann, Rochelle L. and Rothbart, Scott B. and Chandler, Chelsea and Shen, Hui and Coffman, Lan G.},
- title={Epigenomic Reprogramming toward Mesenchymal-Epithelial Transition in Ovarian-Cancer-Associated Mesenchymal Stem Cells Drives Metastasis},
- journal={Cell Rep.},
- volume={33},
- number={10},
- issn={2211-1247},
- doi={10.1016/j.celrep.2020.108473},
- url={https://doi.org/10.1016/j.celrep.2020.108473},
+}
2020
NLP
Common Sense or World Knowledge? Investigating Adapter-Based Knowledge Injection into Pretrained Transformers
@article{lauscher2020common,
+ title={Common Sense or World Knowledge? Investigating Adapter-Based Knowledge Injection into Pretrained Transformers},
+ author={Lauscher, Anne and Majewska, Olga and Ribeiro, Leonardo FR and Gurevych, Iryna and Rozanov, Nikolai and Glava{\v{s}}, Goran},
+ journal={arXiv preprint arXiv:2005.11787},year={2020}
-}
2019
Am. J. Clin. Pathol.
Long Interspersed Nuclear Element 1 Retrotransposons Become Deregulated during the Development of Ovarian Cancer Precursor Lesions
Pisanic, Thomas R., Asaka, Shiho, Lin, Shiou-Fu, Yen, Ting-Tai, Sun, Hanru, Bahadirli-Talbott, Asli, Wang, Tza-Huei, Burns, Kathleen H., Wang, Tian-Li, and Shih, Ie-Ming
There is growing evidence that most high-grade serous ovarian carcinomas likely arise from local dissemination of precursor lesions of the fallopian tube. Evolution of these lesions from early p53 signatures to latter-stage, serous tubal intraepithelial carcinomas (STICs) is characterized by cytologic atypia, accumulation of somatic mutations, and genomic instability, the etiologies of which remain unclear. Long interspersed element 1 (LINE-1) retrotransposon is expressed in many carcinomas, including high-grade serous ovarian carcinoma, where it contributes to genomic instability; however, the timing of LINE-1 activation during this evolution has yet to be elucidated. In this study, we assessed LINE-1 open reading frame 1 protein expression in 12 p53 signature lesions, 32 STICs, and 112 various types of ovarian cancers via immunohistochemical staining and examined LINE-1 promoter methylation in representative cases. We found that 78% and 57% of STICs, with and without concurrent ovarian carcinomas, respectively, exhibited intense LINE-1 immunoreactivity compared with adjacent, normal-appearing fallopian tube epithelium. Hypomethylation of the LINE-1 promoter was found in all STICs exhibiting overexpression. None of the 12 p53 signatures demonstrated significant LINE-1 expression. In ovarian cancer, 84 (75%) of 112 ovarian carcinomas overexpressed LINE-1. Our results indicate that LINE-1 retrotransposons often become deregulated during progression of ovarian cancer precursor lesions from the p53 signature to STIC stages and remain highly expressed in carcinoma.
@article{RN37,
- author={Pisanic, Thomas R. and Asaka, Shiho and Lin, Shiou-Fu and Yen, Ting-Tai and Sun, Hanru and Bahadirli-Talbott, Asli and Wang, Tza-Huei and Burns, Kathleen H. and Wang, Tian-Li and Shih, Ie-Ming},
- title={Long Interspersed Nuclear Element 1 Retrotransposons Become Deregulated during the Development of Ovarian Cancer Precursor Lesions},
- journal={The American Journal of Pathology},
- volume={189},
- number={3},
- pages={513-520},
- issn={0002-9440},
- doi={https://doi.org/10.1016/j.ajpath.2018.11.005},
- url={http://www.sciencedirect.com/science/article/pii/S0002944018306904},
+}
2019
NLP
LIDA: Lightweight Interactive Dialogue Annotator
Collins, Edward, Rozanov, Nikolai, and Zhang, Bingbing
In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP): System Demonstrations, 2019
@inproceedings{collins2019lida,
+ title={LIDA: Lightweight Interactive Dialogue Annotator},
+ author={Collins, Edward and Rozanov, Nikolai and Zhang, Bingbing},
+ booktitle={Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP): System Demonstrations},
+ pages={121--126},year={2019}
-}
Lab Chip
Multilayer microfluidic array for highly efficient sample loading and digital melt analysis of DNA methylation
O’Keefe, Christine M., Giammanco, Daniel, Li, Sixuan, Pisanic, Thomas R., and Wang, Tza-Huei Jeff
Liquid biopsies contain a treasure of genetic and epigenetic biomarkers that contain information for the detection and monitoring of human disease. DNA methylation is an epigenetic modification that is critical to determining cellular phenotype and often becomes altered in many disease states. In cancer, aberrant DNA methylation contributes to carcinogenesis and can profoundly affect tumor evolution, metastatic potential, and resistance to therapeutic intervention. However, current technologies are not well-suited for quantitative assessment of DNA methylation heterogeneity, especially in challenging samples such as liquid biopsies with low DNA input and high background. We present a multilayer microfluidic device for quantitative analysis of DNA methylation by digital PCR and high resolution melt (HRM). The multilayer design facilitates high-density array digitization aimed at maximizing sample loading efficiency. The platform achieves highly parallelized digital PCR-HRM-based discrimination of rare heterogeneous DNA methylation as low as 0.0001% methylated/unmethylated molecules of a classic tumor suppressor gene, CDKN2A (p14ARF).
@article{RN28,
- author={O'Keefe, Christine M. and Giammanco, Daniel and Li, Sixuan and Pisanic, Thomas R. and Wang, Tza-Huei Jeff},
- title={Multilayer microfluidic array for highly efficient sample loading and digital melt analysis of DNA methylation},
- journal={Lab on a Chip},
- volume={19},
- number={3},
- pages={444-451},
- issn={1473-0197},
- doi={10.1039/C8LC01189C},
- url={http://dx.doi.org/10.1039/C8LC01189C},
- year={2019}
-}
Clin. Epigenetics
Promoter methylation of ADAMTS1 and BNC1 as potential biomarkers for early detection of pancreatic cancer in blood
Eissa, Maryam A. L., Lerner, Lane, Abdelfatah, Eihab, Shankar, Nakul, Canner, Joseph K., Hasan, Nesrin M., Yaghoobi, Vesal, Huang, Barry, Kerner, Zachary, Takaesu, Felipe, Wolfgang, Christopher, Kwak, Ruby, Ruiz, Michael, Tam, Matthew, Pisanic, Thomas R., Iacobuzio-Donahue, Christine A., Hruban, Ralph H., He, Jin, Wang, Tza-Huei, Wood, Laura D., Sharma, Anup, and Ahuja, Nita
Despite improvements in cancer management, most pancreatic cancers are still diagnosed at an advanced stage. We have recently identified promoter DNA methylation of the genes ADAMTS1 and BNC1 as potential blood biomarkers of pancreas cancer. In this study, we validate this biomarker panel in peripheral cell-free tumor DNA of patients with pancreatic cancer.
@article{RN38,
- author={Eissa, Maryam A. L. and Lerner, Lane and Abdelfatah, Eihab and Shankar, Nakul and Canner, Joseph K. and Hasan, Nesrin M. and Yaghoobi, Vesal and Huang, Barry and Kerner, Zachary and Takaesu, Felipe and Wolfgang, Christopher and Kwak, Ruby and Ruiz, Michael and Tam, Matthew and Pisanic, Thomas R. and Iacobuzio-Donahue, Christine A. and Hruban, Ralph H. and He, Jin and Wang, Tza-Huei and Wood, Laura D. and Sharma, Anup and Ahuja, Nita},
- title={Promoter methylation of ADAMTS1 and BNC1 as potential biomarkers for early detection of pancreatic cancer in blood},
- journal={Clinical Epigenetics},
- volume={11},
- number={1},
- pages={59},
- issn={1868-7083},
- doi={10.1186/s13148-019-0650-0},
- url={https://doi.org/10.1186/s13148-019-0650-0},
- year={2019}
-}
EBioMedicine
Rab8 GTPase regulates Klotho-mediated inhibition of cell growth and progression by directly modulating its surface expression in human non-small cell lung cancer
Chen, Bo, Huang, Shuhong, Pisanic, Thomas R., Stark, Alejandro, Tao, Yong, Cheng, Bei, Li, Yue, Wei, Yunyan, Zhao, Weihong, Wang, Tza-Huei, and Wu, Jianqing
@article{RN41,
- author={Chen, Bo and Huang, Shuhong and Pisanic, Thomas R. and Stark, Alejandro and Tao, Yong and Cheng, Bei and Li, Yue and Wei, Yunyan and Zhao, Weihong and Wang, Tza-Huei and Wu, Jianqing},
- title={Rab8 GTPase regulates Klotho-mediated inhibition of cell growth and progression by directly modulating its surface expression in human non-small cell lung cancer},
- journal={EBioMedicine},
- volume={49},
- pages={118-132},
- issn={2352-3964},
- doi={10.1016/j.ebiom.2019.10.040},
- url={https://doi.org/10.1016/j.ebiom.2019.10.040},
- year={2019}
-}
2018
Clin. Cancer Res.
Methylomic Analysis of Ovarian Cancers Identifies Tumor-Specific Alterations Readily Detectable in Early Precursor Lesions
Pisanic, Thomas R., Cope, Leslie M., Lin, Shiou-Fu, Yen, Ting-Tai, Athamanolap, Pornpat, Asaka, Ryoichi, Nakayama, Kentaro, Fader, Amanda N., Wang, Tza-Huei, Shih, Ie-Ming, and Wang, Tian-Li
Purpose: High-grade serous ovarian carcinoma (HGSOC) typically remains undiagnosed until advanced stages when peritoneal dissemination has already occurred. Here, we sought to identify HGSOC-specific alterations in DNA methylation and assess their potential to provide sensitive and specific detection of HGSOC at its earliest stages.Experimental Design: MethylationEPIC genome-wide methylation analysis was performed on a discovery cohort comprising 23 HGSOC, 37 non-HGSOC malignant, and 36 histologically unremarkable gynecologic tissue samples. The resulting data were processed using selective bioinformatic criteria to identify regions of high-confidence HGSOC-specific differential methylation. Quantitative methylation-specific real-time PCR (qMSP) assays were then developed for 8 of the top-performing regions and analytically validated in a cohort of 90 tissue samples. Lastly, qMSP assays were used to assess and compare methylation in 30 laser-capture microdissected (LCM) fallopian tube epithelia samples obtained from cancer-free and serous tubal intraepithelial carcinoma (STIC) positive women.Results: Bioinformatic selection identified 91 regions of robust, HGSOC-specific hypermethylation, 23 of which exhibited an area under the receiver-operator curve (AUC) value ≥ 0.9 in the discovery cohort. Seven of 8 top-performing regions demonstrated AUC values between 0.838 and 0.968 when analytically validated by qMSP in a 90-patient cohort. A panel of the 3 top-performing genes (c17orf64, IRX2, and TUBB6) was able to perfectly discriminate HGSOC (AUC 1.0). Hypermethylation within these loci was found exclusively in LCM fallopian tube epithelia from women with STIC lesions, but not in cancer-free fallopian tubes.Conclusions: A panel of methylation biomarkers can be used to accurately identify HGSOC, even at precursor stages of the disease.
@article{RN30,
- author={Pisanic, Thomas R. and Cope, Leslie M. and Lin, Shiou-Fu and Yen, Ting-Tai and Athamanolap, Pornpat and Asaka, Ryoichi and Nakayama, Kentaro and Fader, Amanda N. and Wang, Tza-Huei and Shih, Ie-Ming and Wang, Tian-Li},
- title={Methylomic Analysis of Ovarian Cancers Identifies Tumor-Specific Alterations Readily Detectable in Early Precursor Lesions},
- journal={Clinical Cancer Research},
- volume={24},
- number={24},
- pages={6536-6547},
- doi={10.1158/1078-0432.Ccr-18-1199},
- year={2018}
-}
Ann. Oncol.
Phase II study of nab-paclitaxel in refractory small bowel adenocarcinoma and CpG island methylator phenotype (CIMP)-high colorectal cancer
Overman, M. J., Adam, L., Raghav, K., Wang, J., Kee, B., Fogelman, D., Eng, C., Vilar, E., Shroff, R., Dasari, A., Wolff, R., Morris, J., Karunasena, E., Pisanic, T., Azad, N., and Kopetz, S.
BackgroundHypermethylation of promoter CpG islands [CpG island methylator phenotype (CIMP)] represents a unique pathway for the development of colorectal cancer (CRC), characterized by lack of chromosomal instability and a low rate of adenomatous polyposis coli (APC) mutations, which have both been correlated with taxane resistance. Similarly, small bowel adenocarcinoma (SBA), a rare tumor, also has a low rate of APC mutations. This phase II study evaluated taxane sensitivity in SBA and CIMP-high CRC.Patients and methodsThe primary objective was Response Evaluation Criteria in Solid Tumors version 1.1 response rate. Eligibility included Eastern Cooperative Oncology Group performance status 0/1, refractory disease, and SBA or CIMP-high metastatic CRC. Nab-paclitaxel was initially administered at a dose of 260 mg/m2 every 3 weeks but was reduced to 220 mg/m2 owing to toxicity.ResultsA total of 21 patients with CIMP-high CRC and 13 with SBA were enrolled from November 2012 to October 2014. The efficacy-assessable population (patients who received at least three doses of the treatment) comprised 15 CIMP-high CRC patients and 10 SBA patients. Common grade 3 or 4 toxicities were fatigue (12%), neutropenia (9%), febrile neutropenia (9%), dehydration (6%), and thrombocytopenia (6%). No responses were seen in the CIMP-high CRC cohort and two partial responses were seen in the SBA cohort. Median progression-free survival was significantly greater in the SBA cohort than in the CIMP-high CRC cohort (3.2 months compared with 2.1 months, P = 0.03). Neither APC mutation status nor CHFR methylation status correlated with efficacy in the CIMP-high CRC cohort. In vivo testing of paclitaxel in an SBA patient-derived xenograft validated the activity of taxanes in this disease type.ConclusionAlthough preclinical studies suggested taxane sensitivity was associated with chromosomal stability and wild-type APC, we found that nab-paclitaxel was inactive in CIMP-high metastatic CRC. Nab-paclitaxel may represent a novel therapeutic option for SBA.
@article{RN16,
- author={Overman, M. J. and Adam, L. and Raghav, K. and Wang, J. and Kee, B. and Fogelman, D. and Eng, C. and Vilar, E. and Shroff, R. and Dasari, A. and Wolff, R. and Morris, J. and Karunasena, E. and Pisanic, T. and Azad, N. and Kopetz, S.},
- title={Phase II study of nab-paclitaxel in refractory small bowel adenocarcinoma and CpG island methylator phenotype (CIMP)-high colorectal cancer},
- journal={Annals of Oncology},
- volume={29},
- number={1},
- pages={139-144},
- issn={0923-7534},
- doi={10.1093/annonc/mdx688},
- url={http://dx.doi.org/10.1093/annonc/mdx688},
+}
2018
NLP
Evolutionary Data Measures: Understanding the Difficulty of Text Classification Tasks
Collins, Edward, Rozanov, Nikolai, and Zhang, Bingbing
In Proceedings of the 22nd Conference on Computational Natural Language Learning, 2018
@inproceedings{collins2018evolutionary,
+ title={Evolutionary Data Measures: Understanding the Difficulty of Text Classification Tasks},
+ author={Collins, Edward and Rozanov, Nikolai and Zhang, Bingbing},
+ booktitle={Proceedings of the 22nd Conference on Computational Natural Language Learning},
+ pages={380--391},year={2018}
-}
Sci. Adv.
Facile profiling of molecular heterogeneity by microfluidic digital melt
O’Keefe, Christine M., Pisanic, Thomas R., Zec, Helena, Overman, Michael J., Herman, James G., and Wang, Tza-Huei
This work presents a digital microfluidic platform called HYPER-Melt (high-density profiling and enumeration by melt) for highly parallelized copy-by-copy DNA molecular profiling. HYPER-Melt provides a facile means of detecting and assessing sequence variations of thousands of individual DNA molecules through digitization in a nanowell microchip array, allowing amplification and interrogation of individual template molecules by detecting HRM fluorescence changes due to sequence-dependent denaturation. As a model application, HYPER-Melt is used here for the detection and assessment of intermolecular heterogeneity of DNA methylation within the promoters of classical tumor suppressor genes. The capabilities of this platform are validated through serial dilutions of mixed epialleles, with demonstrated detection limits as low as 1 methylated variant in 2 million unmethylated templates (0.00005%) of a classic tumor suppressor gene, CDKN2A (p14ARF). The clinical potential of the platform is demonstrated using a digital assay for NDRG4, a tumor suppressor gene that is commonly methylated in colorectal cancer, in liquid biopsies of healthy and colorectal cancer patients. Overall, the platform provides the depth of information, simplicity of use, and single-molecule sensitivity necessary for rapid assessment of intermolecular variation contributing to genetic and epigenetic heterogeneity for challenging applications in embryogenesis, carcinogenesis, and rare biomarker detection.
@article{RN39,
- author={O’Keefe, Christine M. and Pisanic, Thomas R. and Zec, Helena and Overman, Michael J. and Herman, James G. and Wang, Tza-Huei},
- title={Facile profiling of molecular heterogeneity by microfluidic digital melt},
- journal={Science Advances},
- volume={4},
- number={9},
- pages={eaat6459},
- doi={10.1126/sciadv.aat6459},
- year={2018},
-}
2017
Semin. Cell Dev. Biol.
Defining, distinguishing and detecting the contribution of heterogeneous methylation to cancer heterogeneity
DNA methylation is a fundamental means of epigenetic gene regulation that occurs in virtually all cell types. In many higher organisms, including humans, it plays vital roles in cell differentiation and homeostatic maintenance of cell phenotype. The control of DNA methylation has traditionally been attributed to a highly coordinated, linear process, whose dysregulation has been associated with numerous pathologies including cancer, where it occurs early in, and even prior to, the development of neoplastic tissues. Recent experimental evidence has demonstrated that, contrary to prevailing paradigms, methylation patterns are actually maintained through inexact, dynamic processes. These processes normally result in minor stochastic differences between cells that accumulate with age. However, various factors, including cancer itself, can lead to substantial differences in intercellular methylation patterns, viz. methylation heterogeneity. Advancements in molecular biology techniques are just now beginning to allow insight into how this heterogeneity contributes to clonal evolution and overall cancer heterogeneity. In the current review, we begin by presenting a didactic overview of how the basal bimodal methylome is established and maintained. We then provide a synopsis of some of the factors that lead to the accrual of heterogeneous methylation and how this heterogeneity may lead to gene silencing and impact the development of cancerous phenotypes. Lastly, we highlight currently available methylation assessment techniques and discuss their suitability to the study of heterogeneous methylation. (C) 2016 Published by Elsevier Ltd.
@article{RN1,
- author={Pisanic, T. R. and Athamanolap, P. and Wang, T. H.},
- title={Defining, distinguishing and detecting the contribution of heterogeneous methylation to cancer heterogeneity},
- journal={Seminars in Cell & Developmental Biology},
- volume={64},
- pages={5-17},
- issn={1084-9521},
- doi={10.1016/j.semcdb.2016.08.030},
- url={<Go to ISI>://WOS:000398612700002},
- year={2017}
-}
Cancer Res.
Abstract 4666: DREAMing as a simple and low cost alternative for the assessment of methylation in ultra rare DNA
Pisanic, Thomas R., Athamanolap, Pornpat, Miller, Brendan F., Wu, Vincent, Elnitski, Laura, and Wang, Tza-Huei
Proceedings: AACR Annual Meeting 2017; April 1-5, 2017; Washington, DCBackground: Current approaches for the assessment of methylation, such as methylation-specific PCR (MSP) and next-generation bisulfite sequencing (BS-Seq) are fundamentally limited in their ability to detect and assess heterogeneous methylation patterns (epialleles) in ultra-rare (<0.1%) DNA. These limitations critically compromise diagnostic utility and render them ill suited for many emerging applications in cancer diagnostics, such as the analysis of methylation heterogeneity in cell-free DNA (cfDNA) and rare cell populations. We recently addressed the need for a low cost alternative to the assessment of methylation of ultra-rare DNA with the development of DREAMing (Discrimination of Rare EpiAlleles by Melt), which uses semi-limiting dilution and precise melt curve analysis to distinguish and enumerate individual copies of DNA at single copy sensitivity and single-CpG-site resolution. Here, we seek to demonstrate the advantages of the DREAMing method over conventional approaches to methylation assessment.Methods: We expand upon the underlying theory of DREAMing and provide guidelines for the development of single-copy sensitive DREAMing assays. We further elucidate methods for tailoring DREAMing assays to samples of interest and compare the performance of these assays to commonly employed techniques including quantitative MSP (qMSP) and BS-Seq.Results: Development of single-copy sensitive DREAMing assays for a number of loci associated with classic tumor-specific methylation such as CHFR and RASSF1A as well as a candidate pan-cancer locus are reported. These assays are then used to analyze methylation in cfDNA derived from the plasma of cancer-positive and healthy patients. DREAM analysis reveals that DREAMing can readily detect over an order of magnitude more epialleles when directly compared to qMSP and BS-Seq assays of the same locus. Some of the challenges associated with distinguishing potential tumor-specific aberrant methylation from background methylation are then discussed and proposed solutions are demonstrated. Lastly, methods for optimizing DREAMing assays for specific sample types are discussed.Conclusions: DREAMing is a recently introduced method for the assessment of locus-specific methylation in samples containing ultra rare target DNA. Its low cost and simplicity coupled with the ability to provide enhanced, single-copy detection of heterogeneous methylation make DREAMing an attractive option over traditional techniques for demanding specimens such as cfDNA and rare cell populations. DREAMing has potential utility in the evaluation of DNA methylation dynamics in cell populations, prenatal testing, as well as clear use in early cancer diagnostic, companion diagnostic and predictive applications.Citation Format: Thomas R. Pisanic, Pornpat Athamanolap, Brendan F. Miller, Vincent Wu, Laura Elnitski, Tza-Huei Wang. DREAMing as a simple and low cost alternative for the assessment of methylation in ultra rare DNA [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4666. doi:10.1158/1538-7445.AM2017-4666
@article{RN33,
- author={Pisanic, Thomas R. and Athamanolap, Pornpat and Miller, Brendan F. and Wu, Vincent and Elnitski, Laura and Wang, Tza-Huei},
- title={Abstract 4666: DREAMing as a simple and low cost alternative for the assessment of methylation in ultra rare DNA},
- journal={Cancer Research},
- volume={77},
- number={13 Supplement},
- pages={4666-4666},
- doi={10.1158/1538-7445.Am2017-4666},
- year={2017}
-}
2016
Biomed. Microdevices
A parallelized microfluidic DNA bisulfite conversion module for streamlined methylation analysis
Aberrant methylation of DNA has been identified as an epigenetic biomarker for numerous cancer types. The vast majority of techniques aimed at detecting methylation require bisulfite conversion of the DNA sample prior to analysis, which until now has been a benchtop process. Although microfluidics has potential benefits of simplified operation, sample and reagent economy, and scalability, bisulfite conversion has yet to be implemented in this format. Here, we present a novel droplet microfluidic design that facilitates rapid bisulfite conversion by reducing the necessary processing steps while retaining comparable performance to existing methods. This new format has a reduced overall processing time and is readily scalable for use in high throughput DNA methylation analysis.
@article{RN34,
- author={Stark, Alejandro and Shin, Dong Jin and Pisanic, Thomas and Hsieh, Kuangwen and Wang, Tza-Huei},
- title={A parallelized microfluidic DNA bisulfite conversion module for streamlined methylation analysis},
- journal={Biomedical Microdevices},
- volume={18},
- number={1},
- pages={5},
- issn={1572-8781},
- doi={10.1007/s10544-015-0029-8},
- url={https://doi.org/10.1007/s10544-015-0029-8},
- year={2016}
-}
2015
Nucleic Acids Res.
DREAMing: a simple and ultrasensitive method for assessing intratumor epigenetic heterogeneity directly from liquid biopsies
Pisanic II, Thomas R., Athamanolap, Pornpat, Poh, Weijie, Chen, Chen, Hulbert, Alicia, Brock, Malcolm V., Herman, James G., and Wang, Tza-Huei
Many cancers comprise heterogeneous populations of cells at primary and metastatic sites throughout the body. The presence or emergence of distinct subclones with drug-resistant genetic and epigenetic phenotypes within these populations can greatly complicate therapeutic intervention. Liquid biopsies of peripheral blood from cancer patients have been suggested as an ideal means of sampling intratumor genetic and epigenetic heterogeneity for diagnostics, monitoring and therapeutic guidance. However, current molecular diagnostic and sequencing methods are not well suited to the routine assessment of epigenetic heterogeneity in difficult samples such as liquid biopsies that contain intrinsically low fractional concentrations of circulating tumor DNA (ctDNA) and rare epigenetic subclonal populations. Here we report an alternative approach, deemed DREAMing (Discrimination of Rare EpiAlleles by Melt), which uses semi-limiting dilution and precise melt curve analysis to distinguish and enumerate individual copies of epiallelic species at single-CpG-site resolution in fractions as low as 0.005%, providing facile and inexpensive ultrasensitive assessment of locus-specific epigenetic heterogeneity directly from liquid biopsies. The technique is demonstrated here for the evaluation of epigenetic heterogeneity at p14ARF and BRCA1 gene-promoter loci in liquid biopsies obtained from patients in association with non-small cell lung cancer (NSCLC) and myelodysplastic/myeloproliferative neoplasms (MDS/MPN), respectively.
@article{RN75,
- author={Pisanic II, Thomas R. and Athamanolap, Pornpat and Poh, Weijie and Chen, Chen and Hulbert, Alicia and Brock, Malcolm V. and Herman, James G. and Wang, Tza-Huei},
- title={DREAMing: a simple and ultrasensitive method for assessing intratumor epigenetic heterogeneity directly from liquid biopsies},
- journal={Nucleic Acids Research},
- volume={43},
- number={22},
- pages={e154-e154},
- issn={0305-1048},
- doi={10.1093/nar/gkv795},
- url={https://doi.org/10.1093/nar/gkv795},
- year={2015},
-}
2014
Analyst
Quantum dots in diagnostics and detection: principles and paradigms
Quantum dots are semiconductor nanocrystals that exhibit exceptional optical and electrical behaviors not found in their bulk counterparts. Following seminal work in the development of water-soluble quantum dots in the late 1990’ s, researchers have sought to develop interesting and novel ways of exploiting the extraordinary properties of quantum dots for biomedical applications. Since that time, over 10 000 articles have been published related to the use of quantum dots in biomedicine, many of which regard their use in detection and diagnostic bioassays. This review presents a didactic overview of fundamental physical phenomena associated with quantum dots and paradigm examples of how these phenomena can and have been readily exploited for manifold uses in nanobiotechnology with a specific focus on their implementation in in vitro diagnostic assays and biodetection.
@article{RN5,
- author={Pisanic, T. R. and Zhang, Y. and Wang, T. H.},
- title={Quantum dots in diagnostics and detection: principles and paradigms},
- journal={Analyst},
- volume={139},
- number={12},
- pages={2968-2981},
- issn={0003-2654},
- doi={10.1039/c4an00294f},
- url={<Go to ISI>://WOS:000336737900003},
- year={2014}
-}
Expert Rev. Mol. Diagn.
The promise of methylation on beads for cancer detection and treatment
Guzzetta, A. A., Pisanic, T. R., Sharma, P., Yi, J. M., Stark, A., Wang, T. H., and Ahuja, N.
Despite numerous technical hurdles, the realization of true personalized medicine is becoming a progressive reality for the future of patient care. With the development of new techniques and tools to measure the genetic signature of tumors, biomarkers are increasingly being used to detect occult tumors, determine the choice of treatment and predict outcomes. Methylation of CpG islands at the promoter region of genes is a particularly exciting biomarker as it is cancer-specific. Older methods to detect methylation were cumbersome, operator-dependent and required large amounts of DNA. However, a newer technique called methylation on beads has resulted in a more uniform, streamlined and efficient assay. Furthermore, methylation on beads permits the extraction and processing of miniscule amounts of methylated tumor DNA in the peripheral blood. Such a technique may aid in the clinical detection and treatment of cancers in the future.
@article{RN4,
- author={Guzzetta, A. A. and Pisanic, T. R. and Sharma, P. and Yi, J. M. and Stark, A. and Wang, T. H. and Ahuja, N.},
- title={The promise of methylation on beads for cancer detection and treatment},
- journal={Expert Review of Molecular Diagnostics},
- volume={14},
- number={7},
- pages={845-852},
- issn={1473-7159},
- doi={10.1586/14737159.2014.943665},
- url={<Go to ISI>://WOS:000340829200007},
- year={2014}
-}
2013
Nanomedicine
In vivo nanoneurotoxicity screening using oxidative stress and neuroinflammation paradigms
Kim, Y., Kong, S. D., Chen, L. H., Pisanic, T. R., Jin, S., and Shubayev, V. I.
Nanomedicine-Nanotechnology Biology and Medicine, 2013
Iron oxide nanoparticles (IONPs) are promising neuroimaging agents and molecular cargo across neurovascular barriers. Development of intrinsically safe IONP chemistries requires a robust in vivo nanoneurotoxicity screening model. Herein, we engineered four IONPs of different surface and core chemistries: DMSA-Fe2O3, DMSA-Fe3O4, PEG-Fe3O4 and PEG-Au-Fe3O4. Capitalizing on the ability of the peripheral nervous system to recruit potent immune cells from circulation, we characterized a spatiotemporally controlled platform for the study of in vivo nanobiointerfaces with hematogenous immune cells, neuroglial and neurovascular units after intraneural IONP delivery into rat sciatic nerve. SQUID magnetometry and histological iron stain were used for IONP tracking. Among the IONPs, DMSA-Fe2O3 NPs were potent pro-apoptotic agents in nerve, with differential ability to regulate oxidative stress, inflammation and apoptotic signaling in neuroglia, macrophages, lymphocytes and endothelial cells. This platform aims to facilitate the development of predictive paradigms of nanoneurotoxicity based on mechanistic investigation of relevant in vivo bio-nanointerfaces. Published by Elsevier Inc.
@article{RN7,
- author={Kim, Y. and Kong, S. D. and Chen, L. H. and Pisanic, T. R. and Jin, S. and Shubayev, V. I.},
- title={In vivo nanoneurotoxicity screening using oxidative stress and neuroinflammation paradigms},
- journal={Nanomedicine-Nanotechnology Biology and Medicine},
- volume={9},
- number={7},
- pages={1057-1066},
- issn={1549-9634},
- doi={10.1016/j.nano.2013.05.002},
- url={<Go to ISI>://WOS:000324520900023},
- year={2013}
-}
Clin. Chim. Acta
Extraction and processing of circulating DNA from large sample volumes using methylation on beads for the detection of rare epigenetic events
Keeley, B., Stark, A., Pisanic, T. R., Kwak, R., Zhang, Y., Wrangle, J., Baylin, S., Herman, J., Ahuja, N., Brock, M. V., and Wang, T. H.
The use of methylated tumor-specific circulating DNA has shown great promise as a potential cancer biomarker. Nonetheless, the relative scarcity of tumor-specific circulating DNA presents a challenge for traditional DNA extraction and processing techniques. Here we demonstrate a single tube extraction and processing technique dubbed "methylation on beads" that allows for DNA extraction and bisulfite conversion for up to 2 ml of plasma or serum. In comparison to traditional techniques including phenol chloroform and alcohol extraction, methylation on beads yields a 15- to 5-fold improvement in extraction efficiency. The technique results in far less carryover of PCR inhibitors yielding analytical sensitivity improvements of over 25-fold. The combination of improved recovery and sensitivity make possible the detection of rare epigenetic events and the development of high sensitivity epigenetic diagnostic assays. (C) 2013 Elsevier B.V. All rights reserved.
@article{RN6,
- author={Keeley, B. and Stark, A. and Pisanic, T. R. and Kwak, R. and Zhang, Y. and Wrangle, J. and Baylin, S. and Herman, J. and Ahuja, N. and Brock, M. V. and Wang, T. H.},
- title={Extraction and processing of circulating DNA from large sample volumes using methylation on beads for the detection of rare epigenetic events},
- journal={Clinica Chimica Acta},
- volume={425},
- pages={169-175},
- issn={0009-8981},
- doi={10.1016/j.cca.2013.07.023},
- url={<Go to ISI>://WOS:000328306100031},
- year={2013}
-}
\ No newline at end of file
+}
2017
2016
2015
2014
2013
\ No newline at end of file
diff --git a/sitemap.xml b/sitemap.xml
index 67ee98ce2936..50486cc5463f 100644
--- a/sitemap.xml
+++ b/sitemap.xml
@@ -1 +1 @@
- https://hunky-d0ry.github.io/members/marek2024-12-12T03:32:37+00:00https://hunky-d0ry.github.io/members/nikolai2024-12-12T03:32:37+00:00https://hunky-d0ry.github.io/2022/04/23/displaying-external-posts-on-your-al-folio-blog.html2022-04-23T23:20:09+00:00https://hunky-d0ry.github.io/https://hunky-d0ry.github.io/publications/https://hunky-d0ry.github.io/team.htmlhttps://hunky-d0ry.github.io/blog/2022/
\ No newline at end of file
+ https://hunky-d0ry.github.io/members/joe2024-12-13T03:27:28+00:00https://hunky-d0ry.github.io/members/lisa2024-12-13T03:27:28+00:00https://hunky-d0ry.github.io/members/marek2024-12-13T03:27:28+00:00https://hunky-d0ry.github.io/members/nikolai2024-12-13T03:27:28+00:00https://hunky-d0ry.github.io/members/prof2024-12-13T03:27:28+00:00https://hunky-d0ry.github.io/members/tyler2024-12-13T03:27:28+00:00https://hunky-d0ry.github.io/2022/04/23/displaying-external-posts-on-your-al-folio-blog.html2022-04-23T23:20:09+00:00https://hunky-d0ry.github.io/https://hunky-d0ry.github.io/publications/https://hunky-d0ry.github.io/team.htmlhttps://hunky-d0ry.github.io/blog/2022/
\ No newline at end of file
diff --git a/team.html b/team.html
index 3dca7d4a65cc..963d85bbfae7 100644
--- a/team.html
+++ b/team.html
@@ -1 +1 @@
- team | Imperial NLP
Team
NLP Group at Imperial College London Computing Department.
