Probing the sequence constraints of CR9114 CDR H3 and light chain

ENV setup

conda env create -f environment.yml

or create ENV with commands

conda create -n CDRH3 python=3.9
conda activate CDRH3
conda install -c bioconda -c anaconda -c conda-forge \
  pear \
  biopython \
  cutadapt \
  snakemake \
  argparse \
  openpyxl 

Pipeline Usage

1. Download raw read files in fastq format from NIH SRA database BioProject PRJNA976657
2. Run the Snakemake pipeline ./code/pipeline.sh to get the sequence read count:
3. Run the notebook CDRH3_Tite_seq_plot.Rmd using R for KD NLS regression

Analysis of Light chain variants

1. Run python3 code/analyze_LC.py to compute expression and binding scores for light chain variants and plot sequence logos using Logomaker.
2. Run Rscript code/plot_LC.R to plot analysis of light chain variants.

Analysis of CDR H3 variants

1. Run python3 code/process_KD_table.py to process the KD results.
2. Run python3 code/compute_expression_score.py to compute expression scores for CDR H3 variants.
3. Run Rscript code/plot_CDRH3_QC.R to plot the correlation between replicates.
4. Run Rscript code/plot_CDRH3_KD_distrib.R to plot the distribution of expression and binding scores, and analyzed the effect of Y98 on binding.
5. Run Rscript code/plot_CDRH3_exp_vs_KD.R to plot expression scores vs binding scores.
6. Run Rscript code/plot_CDRH3_KD_heatmap.R to plot the heatmap to show the effect of single mutants on binding.
7. Run Rscript code/plot_aa_freq.R to plot the amino acid usage at VL residues 91 and 96