#Toxicophore Prediction for Insensitive Munitions
The aim of this project was to use frequent itemset mining to identify chemical substructures that tend to predict toxicity. These substructures are called toxicophores. In this case, I wanted to predict acute aquatic toxicity in fathead minnows.
I used R to perform our QSAR. Every step that I performed to generate this report (the qsar_analysis_just_the_facts.html file) is in the qsar_analysis_just_the_facts.Rmd file. I ran this in RStudio using R version 3.2.3.
You'll need the following R libraries:
- arules
- ChemmineR
- ChemmineOB
- ggplot2
- arulesViz
- plotrix
- gdata
- caret
If you'd like more information, please contact me: [email protected]
#Notes: This information is in the public domain. The "raw" ground-truth data is in the file: fhm_chems.sml and is read into R on line 55.