dekegel / paralog_essentiality

Analysis of the link between paralog buffering and variable paralog essentiality

Paralog Essentiality

Source code for Paralog buffering contributes to the variable essentiality of genes in cancer cell lines

Cite as: De Kegel & Ryan (2019) Paralog buffering contributes to the variable essentiality of genes in cancer cell lines. PLoS Genet

Data analysis notebooks overview:

All of the main and most of the supplementary figures for the manuscript are generated in these notebooks.

Notebook	Figures	Brief description
01_paralog_multitarget.ipynb	Fig. S1A, S1B	Investigate paralogs affected by multi-targeting sgRNAs
02_paralogs_vs_essentiality.ipynb	Fig. 2, 3	Compare essentiality of paralogs and singletons
03_WGD_vs_SSD.ipynb	Fig. 4	Compare essentiality of WGD vs. SSD paralogs
04_regression_model.ipynb		Compare regression models for essentiality with different paralog features
05_A1_essentiality_vs_A2_expression.ipynb	Fig. 5, S2B	Identify putative SL pairs using essentiality (CERES scores) and gene expression
06_A2_loss_reasons.ipynb		Identify potential explanations (copy number, mutation) for gene loss
07_putative_SL_features.ipynb	Fig. 6, S2A, S2C, S2D	Identify features of putative SL paralog pairs

NOTE: Figures S3 and S4 consist of Figs 2, 3B, 3C and 4B re-drawn using different fitness score thresholds

Data processing notebooks overview:

These notebooks process raw/third party data for use in the analysis.

Notebook	Figures	Brief description
01_process_protein_complexes.ipynb		Create map of Entrez ID to CORUM protein complex ID (protein complex membership)
02_process_ensembl_paralogs		Generate list of protein-coding paralog pairs with min. 20% sequence identity + some feat. annotations
03_generate_gene_summary.ipynb	Fig. S1C	For each gene, summarize essentiality and paralog features
04_process_depmap_expr.ipynb		Filter CCLE gene expression for use in data analysis nb 5

Running CERES overview:

These notebooks + R scripts are used to re-process logfold change data from DepMap CRISPR screens with CERES, to remove multi-targetting sgRNAs. The output from running all scripts in this folder is local_data\processed\depmap19Q1\gene_scores_11_07_19.csv. The notebooks can run in the conda environment, but the R scripts should be run independently (e.g. in R studio). Code to install the necessary Bioconductor packages + CERES is included in the R scripts. The R scripts also require the bowtie and samtools command line tools which are only available for Linux/macOS.

Order	Notebook or R scripts	Brief description
1	01_preprocess_depmap_data	Extract sgRNAs to align and format files for CERES
2	R_scripts/align_guides.R	Align all sgRNAs to the reference genome (hg19), allowing up to two mismatches
3	02_filter_multi_target_guides	Filter out sgRNAs that align to readthrough genes and/or to multiple genes (incl. w/ mismatch)
4	R_scripts/run_CERES.R	Run CERES with the filtered logfold changes data
5	03_ceres_post_processing	Process CERES output to drop genes targeted by too few guides + normalize scores to reference (non-)essential genes
6	04_essential_cutoff.ipynb	Calculate a cutoff fitness score for cell line specific essentiality

To run Jupyter notebooks:

• Obtain 3rd party data - sources are listed in data_sources
  • The folder for running CERES has its own data_sources file with the third party files that are only needed there
  • Note: this is primarily needed for running notebooks in the 0_running_ceres and 1_data_processing folders
• Set path to 3rd party data directory in environment.yml (3rd_party_dir)
  • Note: this can also be set in the activate environment with: conda env config vars set 3rd_party_data=my_dir
• Create conda environment from the environment.yml file: conda env create -f environment.yml
• Activate conda environment: conda activate paralogSL
• Start notebooks: jupyter notebook