casTLE: Cas9 high-Throughput maximum Likelihood Estimator

Reference: Morgens, David W., et al. "Systematic comparison of CRISPR/Cas9 and RNAi screens for essential genes." Nature biotechnology 34.6 (2016): 634-636.

Installation:

• Please make sure anaconda is installed on your machine. If not, please follow the instructions:

  https://docs.conda.io/en/latest/miniconda.html # Link for system-dependent installation of binary 
  cd Downloads # Location where binary (example name Miniconda3-latest-MacOSX-x86_64.sh) is downloaded
  bash Miniconda3-latest-MacOSX-x86_64.sh # Installation of conda
  

• Please run the following in order, in the command line:

  conda create -name py2 python=2.7
  conda activate py2
  conda install -c bioconda bowtie=1.2.2
  conda install -c bioconda htseq
  conda install -c anaconda scipy=0.19.1
  pip install pp
  

• If all the installations went smoothly, you are ready to run the commands from 'Quick Start'.

File Descriptions:

• Scripts/makeIndices.py: Creation of a mapping for all the guides, used by future steps for association to a gene. Not requred if genome-wide alignment is prefered.
• Scripts/makeCounts.py: Align FASTQ and make count file
• Scripts/analyzeCounts.py: Compares count files using casTLE
• Scripts/analyzeCombo.py: Combines data for two screens
• Scripts/plotGenes_RL.py: Plot guide distrubutions for Genes
• Scripts/addCombo.py: Adds permutations to calculate pvalues
• Scripts/plotDist.py: Visualizes count distribution

Demultiplexing:

In order to convert raw data into fastq files, we make use of bcl2fastq, Illumina conversion tool that can be downloaded from here. In order to proceed with using casTLE, the data is required to be in the fastq format. The following script can be used for the conversion:

runName=$1  # path to data directory, e.g., /mnt/lab_data/Sequencing_Data/150312_NS500418_0107_AH5M5HBGXX
runLabel=$2 # path to the location where the output will be saved e.g. /mnt/lab_data/Sequencing_Data/fastq_files/OUT_DIR

seqDir=/mnt/lab_data/Sequencing_Data/2019/$runName
outDir=/mnt/lab_data/Sequencing_Data/fastq_files/2019/$runLabel

bcl2fastq \
    --runfolder-dir $seqDir \
    --output-dir $outDir \
    --minimum-trimmed-read-length 17 \
    --mask-short-adapter-reads 0 \
    --barcode-mismatches 0 \
    --stats-dir ./ \
    --reports-dir ./ \
    --ignore-missing-filter \
    --min-log-level WARNING

Quick Start Example:

• Please activate your conda environment with:

  conda activate py2
  

• Please cd inside the downloaded casTLE directory.

• Inside casTLE, please unzip the directory 'GenRef.zip'

• Creation of an index file:
  Step 1: Please save the index file (in csv format) in the directory 'Align'.
  Step 2: Please run the following command in the command line:
  

  python ./Scripts/makeIndices.py -o ./Align/attardi1_align.csv attardi1_align_short attardi1_align_long -t
  

• Run makeCounts.py: (Output will be saved within the directory './Data/Results')
  

  python Scripts/makeCounts.py ./Sequencing_Data/220318_RVJS_NucKO_KS_colab/left_flank_S14_L00  Results/left_flank attardi1_align_short   
  python Scripts/makeCounts.py ./Sequencing_Data/220318_RVJS_NucKO_KS_colab/right_flank_S15_L00  Results/right_flank attardi1_align_short 
  python Scripts/makeCounts.py ./Sequencing_Data/220318_RVJS_NucKO_KS_colab/t0_S13_L00  Results/t0 attardi1_align_short 
  python Scripts/makeCounts.py ./Sequencing_Data/220318_RVJS_NucKO_KS_colab/lib_S16_L00  Results/lib attardi1_align_short 
  

  Note: Every time makeCounts.py is successfully run, information about alignment quality is reported. For example:

  Mapping reads
  # reads processed: 7820356
  # reads with at least one reported alignment: 6628625 (84.76%)
  # reads that failed to align: 1191731 (15.24%)
  

  Having an alignment percentage above 80% is a quick sanity check. Here, 84.76% is reported (which is good!).

• Run analyzeCounts.py: (Output will be saved within the directory 'Results')
  

  python Scripts/analyzeCounts.py ./Data/Results/t0_counts.csv ./Data/Results/right_flank_counts.csv t0_v_right
  python Scripts/analyzeCounts.py ./Data/Results/t0_counts.csv ./Data/Results/left_flank_counts.csv t0_v_left
  python Scripts/analyzeCounts.py ./Data/Results/lib_counts.csv ./Data/Results/t0_counts.csv lib_v_t0
  

• Run analyzeCombo: (Output will be saved within the directory 'Results')
  

  python Scripts/analyzeCombo.py ./Results/t0_v_right.csv  ./Results/t0_v_left.csv combo_left_right
  

• Plot the guides for a particular gene (We will look at the gene SPTAN1): (Output will be saved within the curent working directory)
  

  python Scripts/plotGenes_RL.py ./Results/t0_v_left.csv SPTAN1
  

• Plot the points (genes) comparing their effect (combo casTLE effect), to their significance (combo casTLE Score).: (Output will be saved within the curent working directory in files volcano.png and )
  

  python ./Scripts/plotVolcano.py ./Results/combo_left_right.csv
  

• Run addCombo.py: (Column 'casTLE p-value' will be updated in the combo file of the 'Results' directory)

  python Scripts/addCombo.py Results/combo_left_right.csv 1000000
  

• Run plotDist.py: (Output will be saved within the directory 'Results')

  python Scripts/plotDist.py test_output ./Data/Results/t0_counts.csv ./Data/Results/right_flank_counts.csv
  

Questions, problems?

Make a github issue 😄. Please be as clear and descriptive as possible. Please feel free to reach out in person: (TODO: Email)