chenkenbio/predpsi-svr

Predicting alternative splicing change caused by genetic mutations

PredPSI-SVR

PredPSI-SVR was designed to predict the change of percent spliced in (delta-PSI or ) caused by genetic variants for the CAGI 5 vex-seq challenge.

• About and : https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3037023/

Send questions and comments to [email protected]

Requirements

• Operation system: Unix/Linux
• Memory: 4GB at least
• Perl in your PATH
• Python 2
• Python 3 (with numpy package installed)
  If you have trouble installing python 3 or numpy, you can try miniconda

  cd ~/Downloads
  wget -c https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
  chmod +x Miniconda3-latest-Linux-x86_64.sh
  ./Miniconda3-latest-Linux-x86_64.sh             #pay attention to path of installation, we'll use the default path "$HOME/miniconda3" in this tutorial 
  source $HOME/miniconda/bin/activate
  pip install numpy

Getting started

Preparation

Note: If you have the following packages installed on your system, you can skip installing them and just edit path in src/init.sh
We put PrePSI-SVR in $HOME directory as default

1. Download PredPSI-SVR,

   cd ~
   git clone https://github.com/chenkenbio/PredPSI-SVR

2. Download ANNOVAR (http://annovar.openbioinformatics.org/en/latest/user-guide/download/), libsvm (https://www.csie.ntu.edu.tw/~cjlin/libsvm), and samtools (http://www.htslib.org/download/). And move them to PredPSI-SVR/tools.

3. Extract packages:

   cd ~/PredPSI-SVR/tools
   tar -xzvf annovar.latest.tar.gz
   tar -xzvf libsvm-3.23.tar.gz
   tar -xjvf samtools-1.9.tar.bz2
   cd libsvm-3.23
   make all
   cd ../samtools-1.9
   make all
   cd ..

4. Download basic annotation databases for ANNOVAR

   cd ~/PredPSI-SVR/tools/annovar        # PredPSI-SVR/tools
   ./annotate_variation.pl -buildver hg19 -downdb -webfrom annovar ensGene ./humandb/

5. Download third-party database SPIDEX from http://www.openbioinformatics.org/annovar/spidex_download_form.php. Move it to ~/PredPSI-SVR/tools/annovar/humandb/ and decompress with unzip:

unzip hg19_spidex.zip        #working directory: PredPSI-SVR/tools/annovar/humandb

6. Download hg19 genome

cd ~/PredPSI-SVR/genome
wget -c http://hgdownload.soe.ucsc.edu/goldenPath/hg19/bigZips/chromFa.tar.gz
cat *.fa > hg19.fasta
$HOME/PredPSI-SVR/tools/samtools-1.9/samtools faidx hg19.fasta

7. Finally, check variables in src/init.sh, edit them to fit your system

Usage

Example:

cd ~/PredPSI-SVR/
## PredPSI-SVR, with "-p" option
./main.sh example/sample.vcf -p example/sample.psi -o example/outdir
## PredPSI-SVR-noPSI, with out "-p"
./main.sh example/sample.vcf -o example/outdir

Result file is example/outdir/OUTPUT.dpsi

Attention:

The PredPSI-SVR will filter the VCF file at first to remove variants in intergenic regions or distant to splice sites (more than 200 bp ). Threrefore sometimes you will find that there are fewer variants in OUTPUT.psi than your input VCF file.

References

PredPSI-SVR/tools/ese3/ese3_mod.py is modified based a script in SilVA package (Paper: https://www.ncbi.nlm.nih.gov/pubmed/23736532, GitHub: https://github.com/buske/silva)

Citation

Chen, K., Lu, Y., Zhao, H., & Yang, Y. (2019). Predicting the change of exon splicing caused by genetic variant using support vector regression. Human mutation, 40(9), 1235–1242. https://doi.org/10.1002/humu.23785

predpsi-svr's Issues

Error with libsvm and utils.py

I installed libsvm 3.32 and got this error with PredPSI-SVM

./main.sh example/sample.vcf -o example/outdir 

 

>>== Check commands availablity ==<< 

/usr/bin/python 

/usr/bin/python3 

/usr/bin/perl 

/usr/bin/samtools 

** ERROR: check variable "svm_scale" in src/init.sh 

** ERROR: check variable "svm_predict" in src/init.sh 

** ANNOVAR annotating...... 

NOTICE: Finished reading 99 lines from VCF file 

NOTICE: A total of 99 locus in VCF file passed QC threshold, representing 99 SNPs (77 transitions and 22 transversions) and 0 indels/substitutions 

NOTICE: Finished writting 0 SNPs (0 transitions and 0 transversions) and 0 indels/substitutions for 1 sample 

NOTICE: The --geneanno operation is set to ON by default 

NOTICE: Reading gene annotation from /data/jess_tmp/fh/PredPSI-SVR/tools/annovar/humandb/hg19_ensGene.txt ... Done with 103433 transcripts (including 38799 without coding sequence annotation) for 47132 unique genes 

NOTICE: Finished gene-based annotation on 0 genetic variants in /data/jess_tmp/fh/PredPSI-SVR/example/outdir/input.avinput 

NOTICE: Output files were written to /data/jess_tmp/fh/PredPSI-SVR/example/outdir/annovar.variant_function, /data/jess_tmp/fh/PredPSI-SVR/example/outdir/annovar.exonic_variant_function 

    DONE 

** Preparing mutation info... 

  - only vcf, finding exon transcript... 

/data/jess_tmp/fh/PredPSI-SVR/src/utils.py:30: SyntaxWarning: "is" with a literal. Did you mean "=="? 

  assert errmsg is '', "Error in file 'utils.py':'get_genome': " + errmsg 

/data/jess_tmp/fh/PredPSI-SVR/src/utils.py:51: SyntaxWarning: "is" with a literal. Did you mean "=="? 

  if len(args) is 0: 

/data/jess_tmp/fh/PredPSI-SVR/src/utils.py:118: SyntaxWarning: "is" with a literal. Did you mean "=="? 

  if self.strand is '-': 

/data/jess_tmp/fh/PredPSI-SVR/src/utils.py:292: SyntaxWarning: "is" with a literal. Did you mean "=="? 

  if strand is '+': 

However, you can fix it by doing

chmod -x svm-predict.c 

chmod -x svm-scale.c 

And change all the "is" in util.py to "=="

This gives

./main.sh example/sample.vcf -o example/outdir 

 

>>== Check commands availablity ==<< 

/usr/bin/python 

/usr/bin/python3 

/usr/bin/perl 

/usr/bin/samtools 

/data/jess_tmp/fh/PredPSI-SVR/tools/libsvm-3.32/svm-scale.c 

/data/jess_tmp/fh/PredPSI-SVR/tools/libsvm-3.32/svm-predict.c 

** ANNOVAR annotating...... 

NOTICE: Finished reading 99 lines from VCF file 

NOTICE: A total of 99 locus in VCF file passed QC threshold, representing 99 SNPs (77 transitions and 22 transversions) and 0 indels/substitutions 

NOTICE: Finished writting 0 SNPs (0 transitions and 0 transversions) and 0 indels/substitutions for 1 sample 

NOTICE: The --geneanno operation is set to ON by default 

NOTICE: Reading gene annotation from /data/jess_tmp/fh/PredPSI-SVR/tools/annovar/humandb/hg19_ensGene.txt ... Done with 103433 transcripts (including 38799 without coding sequence annotation) for 47132 unique genes 

NOTICE: Finished gene-based annotation on 0 genetic variants in /data/jess_tmp/fh/PredPSI-SVR/example/outdir/input.avinput 

NOTICE: Output files were written to /data/jess_tmp/fh/PredPSI-SVR/example/outdir/annovar.variant_function, /data/jess_tmp/fh/PredPSI-SVR/example/outdir/annovar.exonic_variant_function 

    DONE 

** Preparing mutation info... 

  - only vcf, finding exon transcript... 

I haven't got OUTPUT.dpsi but that might be because I haven't got the SPIDEX database installed yet (still waiting for a link)

/PredPSI-SVR/example/outdir$ ls

annovar.exonic_variant_function  annovar.log  annovar.variant_function  input.avinput  input.vcf  mut_info  mut_info.valid