Cancer Mutations Pipeline

Cancer driver identification based on quantitative prediction of the effect of mutations in regulatory elements on transcription factor (TF) binding. The TF binding change predictions are done using QBiC-Pred (https://github.com/vincentiusmartin/QBiC-Pred).

Input dataset:

1. any ICGC dataset. For example: Liver Cancer-RIKEN, Japan; i.e. LIRI-JP, https://dcc.icgc.org/releases/current/Projects/LIRI-JP.
2. Promoter, enhancer, exon data, e.g. from RefSeq. Processed RefSeq data is provided in dataset/epdata folder.

All input paths for the code are defined in each code's header.

1. Read ICGC simple somatic mutations (SSM) and expression data (EXP-S)

Code: readtbl.R
Input:

1. ICGC simple somatic mutation (SSM) data, required fields:

chromosome	start	end	icgc_mutation_id	icgc_donor_id	ref	mut
chr1	565238	565238	MU886122	DO23508	T	C
chr1	569604	569604	MU941619	DO23508	G	A
chr1	570513	570513	MU952816	DO23508	A	G
chr1	754535	754535	MU76676523	DO23508	G	A
chr1	1487203	1487203	MU958359	DO23508	C	A
chr1	1487205	1487205	MU76676613	DO23508	T	C

2. Expression data for the cancer mutations with the overlapping donors (from icgc_donor_id column) with the somatic mutations data. Required fields:

icgc_donor_id	gene_id	normalized_read_count	analysis_id
DO50825	DYNC2LI1	11.654758	RK270_Cancer-rnaseq
DO45221	CTTN	15.685822	RK100_Liver-rnaseq
DO23538	RIC8A	9.196332	RK131_Cancer-rnaseq
DO23545	LOC100505835	0.312690	RK151_Cancer-rnaseq
DO23542	HERPUD2	6.738176	RK143_Liver-rnaseq

3. List of all exons to filter out mutations inside exons

2. Filtering promoter data

Code: filter_coding_promoters.R
Output: dataset/epdata/promoters_refseq_coding.txt

This script takes all promoters and take only promoters for the coding regions. Since promoters region were taken using RefSeq TSS as its center, the script filters for coding regions by looking for TSS with mappings to HGNC genes.

3. Get trinucleotide background dist and mutation rate

Code: trinucleotide_bg.R
Output: dataset/bg_entropy_data

Compute the trinucleotide background mutations for the promoters or enhancers.

4. Getting the entropy

Code: muteffect_entropy.R

This script is used for:

1. Compute the mutation effects on TF binding, set process_type <- "qbic". The input used are prediction files from QBiC-Pred, available to download from http://qbic.genome.duke.edu/downloads.
   For this use case, the script accept command argument as an index for the file in the input directory, this is used for parallel processing using SLURM, see code submit_job.sh.
2. Compute Shannon entropy for the input regulatory element, set process_type <- "entropy". The output files are written to dataset/bg_entropy_data.

5. Annotate enhancer or promoter with its gene mapping

Code: annotate_cisreg.R

The output are regulatory element's coordinates with the genes mapped to each element: dataset/cis_annotated.

6. Combine prediction across regulatory elements

Code: combinepreds.R

Some predictions data are provided, for the complete predictions please contact us.

7. Gene expression analysis of the combined regulatory elements

Code: analyze_combined_preds.R