BayesPI-BAR2 pipeline tools

The BayesPI-BAR2 package is a set of command line tools residing in the python folder. Run python <tool_name.py> --help command to see the full usage information for a particular tool. The purpose of each tool and the data involved in the processing are described below:

• differential_expression.py: computes differentially expressed genes between two groups of samples, usually the patient samples and normal tissue samples. This implements a relatively simple and fast method based on Kolmogorov-Smirnov test comparing two sets of normalized RPKM values. For a more precise analysis, the baySeq R package could be used instead.
  Input: two sets of read count files from RNA-Seq experiments in the tab-delimited format
  Output: list of differentially expressed genes
• gene_regions.py: extracts gene regions based on offsets from transcription start sites.
  Input: gene annotation GTF file
  Output: BED file with selected regions
• intersect_vcf.py: computes intersection between VCF files.
  Input: a set of VCF files
  Output: VCF file
• mussd.py: implements the Mutation filtering based on the Space and Sample Distribution (MuSSD) algorithm which finds DNA regions that are highly mutated in multiple patients (mutation blocks).
  Input: a set of VCF files with patient somatic mutations, one per patient, a BED file with regions of interest, optionally a set of VCF files with patient germline mutations
  Output: a pair of FASTA files for each block, containing reference and alternate sequences for each patient, in the format expected by bayespi_bar.py
• highly_mutated_blocks.py: a simple method to find which blocks appear to have higher than average mutation rate, used to discard blocks with few mutations.
  Input: output of mussd.py
  Output: list of highly mutated blocks
• bayespi_bar.py: implements the BayesPI-BAR [3] method to compute the protein-DNA binding affinity changes due to sequence variants.
  Input: two FASTA files with a set of paired reference and alternate sequences, as well as a folder containing PWM models in BayesPI format. The set of 1772 PWMs used in the pipeline is provided in the data/pwm folder
  Output: rankings of PWMs by predicted effect size, δdbA values for all PWMs and variants
• choose_background_parameters.py: selects appropriate parameters (sequence size and mutation count distribution) for the background model based on chosen foreground (i.e. patient) mutation blocks.
  Input: output of mussd.py and bayespi_bar.py, optionally a mutation signature file or a set of background mutations to use
  Output: Bash script which runs background_affinity_changes.py with chosen parameters
• background_affinity_changes.py: selects random mutation blocks for the background model and computes the background model using bayespi_bar.py.
  Input: BED file with regions of interest, a folder with PWM files, optionally a mutation signature file or a set of background mutations to use
  Output: δdbA values for all PWMs and randomly generated variants (same as bayespi_bar.py)
• affinity_change_significance_test.py: performs the Wilcoxon rank-sum testing, comparing the patient affinity changes in mutation blocks to the background model.
  Input: a pair of outputs of bayespi_bar.py, for the background model and a mutation block
  Output: a table of PWMs which are affected in the foreground block significantly more than expected according to the background model
• plot_result.py: produces heatmaps displaying predicted affinity changes across transcription factors and patient samples.
  Input: output of affinity_change_significance_test.py and bayespi_bar.py
  Output: PNG file