Skip to content

Latest commit

 

History

187 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Somatic evolution of prostate cancer: mutation, selection, and epistasis across disease stages.

Computational analysis of somatic evolution across primary and metastatic castration-resistant prostate cancer (mCRPC) using large-scale, multi-cohort genomic datasets.

This repository contains the R analysis and visualization workflows supporting our study:

Rajaei M., Yang A., Cross C.N., Glasmacher K., Fisk J.N., Perry E.B., Mandell J.D., Gaffney S.G., Yamaguchi T.N., Livingstone J., Costa J., Humphrey P., Cannataro V.L., Boutros P.C., Townsend J.P. Somatic evolution of prostate cancer: mutation, selection, and epistasis across disease stages. bioRxiv (2025).

Project Overview

We analyzed 2,704 primary and mCRPC tumors from multiple cohorts and sequencing platforms to characterize how mutation, selection, and genetic interactions shape prostate cancer evolution.

The study quantifies mutation rates and somatic selection across disease stages and identifies synergistic and antagonistic epistatic interactions among key cancer drivers, revealing a dynamic evolutionary landscape from tumor initiation to mCRPC.

The study combines tumors profiled using:

  • Whole-genome sequencing (WGS)
  • Whole-exome sequencing (WES)
  • Targeted sequencing

Research Questions

The analyses address several major questions:

  1. How do somatic mutation profiles change across stages of prostate cancer?
  2. Which driver genes experience the strongest selection during primary tumor development and metastatic progression?
  3. How do mutation rates and cancer effect sizes differ across disease stages?
  4. Which driver mutations are selected early versus late during tumor evolution?
  5. How does the presence of one driver mutation alter selection on subsequent mutations?
  6. Which genetic interactions may contribute to progression toward metastatic and treatment-resistant disease?

Dataset

The integrated dataset contains 2,704 prostate tumors. Samples were excluded if they contained only nucleotide base substitutions at known germline variant sites, within repetitive regions, or no nucleotide base substitutions, yielding a final set of 2,618 samples for downstream analysis

Disease group Number of tumors
Primary prostate cancer 1,593
Metastatic castration-resistant prostate cancer (mCRPC) 1,025
Total 2,618

Primary tumors were additionally stratified by Gleason Grade Group to investigate evolutionary differences across disease severity.

The combined dataset incorporates multiple independent prostate cancer cohorts, including data from:

  • TCGA
  • MSK cohorts
  • SU2C/PCF
  • Armenia cohort
  • Boutros cohort
  • Additional published prostate cancer sequencing studies

Sequencing Platforms

The combined dataset includes:

Sequencing strategy Approximate number of tumors
Whole-genome sequencing 293
Whole-exome sequencing 1,122
Targeted sequencing 1,289

Analytical Framework

The overall workflow can be summarized as:

Multiple prostate cancer cohorts
              │
              ▼
    Mutation data harmonization
              │
              ▼
   Disease-stage classification
              │
              ▼
  Mutation prevalence analysis
              │
              ▼
Trinucleotide mutation profiles
              │
              ▼
 Gene-specific mutation rates
              │
              ▼
   Cancer effect size analysis
              │
        ┌─────┴─────┐
        ▼           ▼
 Stage-specific   Driver-specific
   selection       selection
        │           │
        └─────┬─────┘
              ▼
    Pairwise epistasis
              │
              ▼
 Evolutionary interpretation

Major Analyses

Mutation Prevalence

Figure_1_prevalence.R

Prevalence of variants in 16 selected driver genes, in low-grade primary tumors, high-grade primary tumors, and mCRPC.

Trinucleotide Mutation Profiles

Figure_2_trinucleotide_mutation_profiles.R

Percent of single-nucleotide somatic variants within each trinucleotide context in low-grade primary tumors, high-grade primary tumors, and mCRPC.

Gene-Specific Mutation Rates

Figure_3_gene_mutrate.R

The gene-level mutation rates spanning from organogenesis to low-grade primary tumors, organogenesis to high-grade primary tumors, organogenesis to mCRPC, and in organogenesis to mCRPC versus organogenesis to low-grade primary tumorigenesis, organogenesis to mCRPC versus organogenesis to high-grade primary tumorigenesis, organogenesis to mCRPC versus primary tumorigenesis.

Cancer Effect Sizes

Figure_4_CES.R

Gene-level estimates and 95% confidence intervals for scaled selection coefficients on somatic variants in oncogenic sites of 16 genes that are known to act as drivers in prostate cancer tumorigenesis and metastasis.

SPOP Evolution

Figure_5_SPOP.R

Scaled selection coefficients for recurrent single-nucleotide variant amino-acid substitutions in SPOP during the evolutionary trajectory from prostate organogenesis to primary and mCRPC tumors.

Related visualization:

SPOP_model_recurrent_resized.png

AR Evolution

Figure_6_AR.R

Scaled selection coefficients of recurrent single-nucleotide variant amino-acid substitutions in AR along the step from primary tumors to mCRPC.

Related visualization:

AR_labeled_2.png

Pairwise Epistatic Effects

Figure_7_Pairwise_Epistatic_Effects.R

Pairwise epistatic effect trends of 15 genes, focusing on SPOP, PIK3CA, TP53, and AR.

Key Biological Findings

  • Mutation load and mutation rates increase during prostate cancer progression, while trinucleotide mutational patterns remain relatively stable.
  • SPOP mutations in the BRD3-binding domain experience strong early positive selection and increase subsequent selection for RHOA while decreasing selection for TP53.
  • CUL3 shows antagonistic selective epistasis with both SPOP and PIK3CA.
  • KMT2C mutations increase selection for subsequent TP53 mutations.
  • PTEN mutations increase selection for both PIK3CA and AR, revealing strong synergistic epistatic interactions.

Together, these results reveal a dynamic landscape of mutation, selection, and epistasis across prostate cancer progression.

Repository Contents

Main Analysis Scripts

File Analysis
Figure_1_prevalence.R Somatic mutation prevalence across disease groups
Figure_2_trinucleotide_mutation_profiles.R Trinucleotide mutation-spectrum analysis
Figure_3_gene_mutrate.R Gene-specific mutation-rate analysis
Figure_4_CES.R Cancer effect size analysis
Figure_5_SPOP.R SPOP-specific evolutionary analysis
Figure_6_AR.R AR-specific evolutionary analysis
Figure_7_Pairwise_Epistatic_Effects.R Pairwise epistasis among prostate cancer drivers
Figure_S1.R Supplementary analysis
Figure_S2.R Supplementary analysis
Figure_S3.R Supplementary analysis
new_sequential_lik.R Sequential likelihood/statistical modeling functions

Mutation Datasets

The repository contains harmonized mutation data from multiple prostate cancer cohorts, including:

MSK_341_final.maf.txt
MSK_410_final.maf.txt
MSK_468_final.maf.txt
SU2C_final.maf.txt
tcga_final.maf.txt
tcga_wgs_final.maf.txt
armenia_final.maf.txt
boutros_final.maf.txt
prad_armenia_final.maf.txt
prad_boutros_wgs_final.maf.txt

Mutation data are primarily represented using Mutation Annotation Format (MAF) files.

Genomic Coverage Files

SureSelect_All_Exon_covered_regions.bed
msk_341_exons.bed
msk_410_exons.bed
msk_468_exons.bed

Clinical and Disease-Stage Information

gleason.txt
gleason_age_comparison.txt

Computational Environment

The analyses were primarily developed using:

  • R
  • Linux
  • High-performance computing environments
  • Git/GitHub for version control

Reproducibility

The main figure scripts correspond directly to analyses presented in the associated manuscript.

A typical analysis workflow is:

1. Obtain and harmonize cohort-specific mutation data
2. Define genomic coverage for each sequencing platform
3. Assign samples to disease-stage groups
4. Analyze mutation prevalence
5. Calculate trinucleotide mutation profiles
6. Estimate gene-specific mutation rates
7. Quantify cancer effect sizes
8. Analyze individual prostate cancer drivers
9. Estimate pairwise epistatic effects
10. Generate manuscript figures

Related Publication

Rajaei M, Yang A, Cross CN, Glasmacher K, Fisk JN, Perry EB, Mandell JD, Gaffney SG, Yamaguchi TN, Livingstone J, Costa J, Humphrey P, Cannataro VL, Boutros PC, Townsend JP. Somatic evolution of prostate cancer: mutation, selection, and epistasis across disease stages. bioRxiv. 2025.

Author

Moein Rajaei, Ph.D. Computational Biologist | Genomics & Bioinformatics Scientist

Citation

If you use code or analytical methods from this repository, please cite:

Rajaei M, et al. Somatic evolution of prostate cancer: mutation, selection, and epistasis across disease stages. bioRxiv. 2025.

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

2 watching

Forks

Releases

Packages

Contributors

Languages