CSVDriver

Bioinformatics tool to identify SV drivers from whole-genome sequences.

A ShyniAppTo of the the method is being implemented to improve usability for the users that are not expert in the R programming language.

To reproduce the result of the study it can be run Reproduce_CSVDriver_Script.R (R version 3.6.2)

The script could run on a desktop computer. Reproduce_CSVDriver_Script.R will perform every step of CSVDriver using the SVs call from a WGS cancer cohort

Running the demo (time 30 min)

Before running the Reproduce_CSVDriver_Script.R with the input demo gennome_annotation.zip file must be decompressed.

• To run the demo just set the cohort_name = "demo"

In the Reproduce_CSVDriver_Script.R cohort_name <- cohort[1]

Then run the complete script to get the full set of results

Input (input folder)

Demo datasets of SV and tissue-specific genomic feature annotations to evaluate the method

• demo.somatic.sv.bedpe

• ChromHMM/demo.bed

• TAD_segment_demo_ChromMark_class.txt

• demo_ccre_enhancers_data_hg19.bed

Output (output folder)

• GAM model GAM_demo

• Peak data info Table_Result_demo_peak_data_info

• Significant gene driver candidate within significant rearranged peak Table_Result_demo_significant_Genes_best_candidate_in_peak

• Gene linked to significant enhancer driver candidate within significant rearranged peak Table_Result_demo_significant_Enhancer_best_candidate_in_peak

• Set of plots for model performance and the contribution of each covariate to the GAM model demo_Model_covariate_contribution.pdf demo_Model_Fitting_and_Residual

• QQ-plot of peaks rearrangement score PRe for significant peaks demo_PRe_QQplot

• Graphic of BPpc demo_BPpc

• Graphic of driver candidates from the BPpc analysis demo_Driver-Cadiadtes

Running in a full SVs dataset from cancer cohort

While runing on a full SV dataset from cancer cohorts the script could take several hours, because the GAM modeling the algorithm is computationally heavy.

• To run CSVDriver in a cohort the SV input files must be placed in the corresponding folders 'input/cancer/[cohort]/' likewise the demo.

Before running the Reproduce_CSVDriver_Script.R, the TAD_segment_[cohort]_ChromMark_class.zip file in the 'input/cancer/[cohort]/' folder must be decompressed for each corresponding cancer cohort.

• To run the a particular cohort set the cohort_name from the list

cohort <- c("demo", "bone", "brain", "breast", "colon", "esophagus", "kidney", "liver", "lung", "lymph-nodes", "ovary", "pancreas", "prostate", "skin", "stomach", "uterus")

cohort_name <- cohort[ cohort_index]

CSVDriver data sources

Genes: GENCODE Release 29 (GRCh37) Comprehensive gene annotation

Enhancer: GENCODE 3 annotations

Replication Timing (RT): ENCODE Repli-seq data; mean of 1 Mb window, average of 8 different tissue type cell line

Fragile sites (FS): HumCFS database

Gene density: Function kpPlotDensity from the R Package kpPlotDensity (compute in 1 Mb windows)

Chromatin mark (ChromMark): Roadmap Epigenomics Mapping Consortium

Repeat Calss (RepClass): reeatmapsker data from UCSC Genome Browser

LAD: Akdemir, K. C. et al. 2020

GC content (GC): Function GCcontent from the R Package biovizBase

TAD: 3D Genome Browser