lilab-bioinfo / cancerapa

Cancer susceptible genes revealed by transcriptome-wide association study

CancerAPA

In this study, we performed the first large-scale and systematic analysis assessing the genetic effects of APA on 22 cancer types in 49 human tissues from the Genotype-Tissue Expression(GTEx v8).

This repository contains all source code for the analyses in manuscript "A distinct class of pan-cancer susceptibility genes revealed by alternative polyadenylation transcriptome-wide association study ".

System requirements

Hardware and operating system requirements

A workstation or computer cluster running a POXIS system (Unix, Linux, or macOS) is required (we used the Linux distribution Ubuntu 18.04.6 LTS). The minimum requirements for the test dataset are an 8-core processor, 16 GB of RAM, and 1 TB of hard disk space.

Software and dependencies

R and dependent packages

• R (>= v3.6.1)
• PEER (v1.3)
• data.table (1.14.2)
• coloc (v5.1.1)
• glmnet (4.1-3)
• MatrixEQTL(v2.1.0)
• RColorBrewer
• optparse (1.7.1)
• plink2R (1.1)

Python and dependent packages

• Python (version >= v2.7.14)
• numpy
• scipy
• argparse

External software

• LDSC (v1.0.1)
• plink 1.9 beta
• CAUSALdb
• tabix (v1.7)
• bedtools (>= v2.30.0)
• samtools (>= v1.10)
• DaPars (v2.1)
• CAVIAR (v2.2)
• fgwas(0.3.6)
• GCTA (1.93.3beta2)
• GEMMA (0.98.5)
• FUSION

Installation guide

We highly recommend using conda to setup and manage the software environment. The instructions below introduce how to setup the running environment with conda.

• Install python and conda. Download installer script “Anaconda2-2019.10-Linux-x86_64.sh” from Anaconda Repository. Python and conda will be installed in $HOME/anaconda2/bin by default.

> bash Anaconda2-2019.10-Linux-x86_64.sh

• Install other software and dependencies with conda

> conda install -c conda-forge r-base
> conda install -c conda-forge r-dplyr
> conda install -c bioconda r-optparse
> conda install -c bioconda Bioconductor-impute
> conda install -c bioconda r-peer
> conda install -c bioconda r-matrixeqtl
> conda install -c bioconda r-coloc
> conda install -c bioconda bedtools
> conda install -c bioconda plink=1.90
> conda install -c bioconda samtools
> conda install -c bioconda vcftools
> conda install -c bioconda tabix

• Install CASUALdb for GWAS summary statistics based fine-mapping

  • Clone the repo:

  > git clone https://github.com/mulinlab/CAUSALdb-finemapping-pip.git
  

  • set up conda environment and download fine-mapping tools:

  > cd bin
  > bash 00_set_up.sh
  > cd ..
  

  • build reference panel:

  > cd ref
  > python 01_prepare_reference.py
  > cd ..
  

• Install LDSC for estimating heritability and genetic correlation from GWAS summary statisctics.

  • Clone the repo:

  > git clone https://github.com/bulik/ldsc.git
  > cd ldsc
  

  • Create an Anaconda environment with LDSC's dependencies:

  > conda env create --file environment.yml
  > source activate ldsc
  

• Install FUSION software package from github:

  • Download and unpack the FUSION software package from github:

  > wget -c  https://github.com/gusevlab/fusion_twas/archive/master.zip
  > unzip master.zip
  > cd fusion_twas-master
  

  • Download and unpack the 1000Genome LD reference data:

  > wget -c https://data.broadinstitute.org/alkesgroup/FUSION/LDREF.tar.bz2
  > tar xjvf LDREF.tar.bz2  
  

  • Download and unpack the plink2R library：

  > wget https://github.com/gabraham/plink2R/archive/master.zip
  > unzip master.zip
  

  • Launch R and install required libraries:

  install.packages(c('optparse','RColorBrewer'))
  install.packages('plink2R-master/plink2R/',repos=NULL)
  

  • the following steps are required for computing weights
    • Add the bundled GCTA binary to path
    • Add plink to path
    • Launch R and install the required libraries：

    install.packages(c('glmnet','methods'))
    

  • If using BSLMM, download and install GEMMA software, add to path. Generate a symbolic link to the output by calling ln -s ./ output in the directory where you will run FUSION.weights.R (this is a workaround because GEMMA requires results to go into an output subdirectory).

Demo data for running the codes

Each analysis has a demo input data under the src/ folder.

Instructions for use

1. Definition of lead SNPs and trait-assocaited loci using plink clumping:

> bash run_plink_clump.sh

2. Fine mapping of trait-associated loci to defined credible SNPs:

> bash run_finemapping.sh

3. Estimate SNP heritability and genetic correlation:

> bash run_ldsc_heritability.sh
> bash run_ldsc_genetic_correlation.sh

4. APA quantification and 3'aQTL mapping:

We have developed a pipeline to analyze APA and call 3'aQTL before. Please refer to our 3aQTL-pipe for detailed instructions.

5. Enrichment of 3'aQTL in cancer GWAS signals

> bash LDSC-Partitioned-Heritability.sh
> bash run-fgwas-cancer.sh  

6. Colocalization of trait-associated loci and 3'aQTL:

> bash run_aQTL_colocalization.sh

7. Build 3'TWAS model and run transcriptome-wide association analysis:

• Compute GTEx v8 APA predictive models:

  Rscript FUSION_compute_weights.R \
  --bfile ${tissName}.${GNAME} \
  --tmp ${tissName}.${GNAME}.tmp \
  --out ${tissName}.${GNAME} \
  --PATH_plink $PLINK \
  --PATH_gcta $GCTA \
  --PATH_gemma $GEMMA \
  --models top1,blup,lasso,enet \
  --covar ${COVAR} \
  --hsq_p 0.05 \
  

• APA-based transcriptome wide association analysis

  • Input: GWAS summary statistics

    The input of GWAS summary statistics is in LD-score format, which mainly contain the following columns:

    1. SNP : SNP identifier(rsID)
    2. A1 : first allele (effect allele)
    3. A2 : second allel (other allele)
    4. Z : Z-score,sign with respect to A1

  • Input: APA weights

    The reference data are loaded using the ./WEIGHTS/${tissueName}.pos which points to the individual *RDat weights files, their gene identifiers, physical positions. Only weights in the file will be evaluated.

  • Performing the APA level-disease association

    Rscript FUSION.assoc_test.R \
    --sumstats $WORK_DIR/GWAS/*.sumstats
    --weights $WORK_DIR/output/WEIGHTS/'${tissueName}'.pos \
    --weights_dir ./WEIGHTS/ \
    --ref_ld_chr ./LDREF/1000G.EUR. \
    --chr ${chr} \
    --out $WORK_DIR/results/'${GWAS}'/'${tissueName}'/'${GWAS}'.'${tissueName}'.chr${chr}.dat 
    

Authors

Hui Chen, Wenyan Chen, Xudong Zou, Lei Li (Institute of Systems and Physical Biology, Shenzhen Bay Laboratory)

Citation

• Hui Chen^, Zeyang Wang^, Jia Wang^, Wenyan Chen, Xuelian Ma, Xudong Zou, Mireya Plass, Cheng Lian, Ting Ni, Gong-Hong Wei, Wei Li#, Lin Deng#, Lei Li# A distinct class of pan-cancer susceptibility genes revealed by alternative polyadenylation transcriptome-wide association study (2023) medrxiv 10.1101/2023.02.28.23286554

• Xudong Zou, Ruofan Ding, Wenyan Chen, Gao Wang, Shumin Cheng, Qin Wang, Wei Li, Lei Li. Using population-scale transcriptomic and genomic data to map 3' UTR alternative polyadenylation quantitative trait loci STAR Protocols,3(3):101566 (2022). DOI: https://doi.org/10.1016/j.xpro.2022.101566

• Lei Li, Kai-Lieh Huang, Yipeng Gao, Ya Cui, Gao Wang, Nathan D. Elrod, Yumei Li, Yiling Elaine Chen, Ping Ji, Fanglue Peng, William K. Russell, Eric J. Wagner & Wei Li.An atlas of alternative polyadenylation quantitative trait loci contributing to complex trait and disease heritability Nature Genetics,53,994-1005 (2021). DOI:https://doi.org/10.1038/s41588-021-00864-5

Contact

If you have any comments, suggestions, questions, etc, please feel free to create a GitHub Issue.

Funding

This work was supported by National Natural Science Foundation of China (no. 32100533) and startup funds from Shenzhen Bay Laboratory to L.L.