Quick Start

JCAST software for alternative splicing proteomics analysis. Create custom protein databases using RNA-seq data to identify unique protein alternative splicing isoforms in mass spectrometry experiments

Abstract

This page provides the documentation for the JCAST package, which is a tool for create a database of protein isoforms from RNA-seq data.

Installing JCAST

Requirements

Install Python 3.7+ and pip. See instructions on Python website for specific instructions for your operating system.

JCAST can be installed from PyPI via pip. We recommend using a virtual environment.

$ pip install jcast

Running JCAST

Launch JCAST as a module (Usage/Help):

$ python -m jcast

Alternatively:

$ jcast

Example command:

$ python -m jcast  data/encode_human_pancreas/ data/gtf/Homo_sapiens.GRCh38.89.gtf data/gtf/Homo_sapiens.GRCh38.89.gtf data/genome/Homo_sapiens.GRCh38.dna.primary_assembly.fa -o encode_human_pancreas -q 0 1 -r 1 -m -c

To test that the installation can load test data files in tests/data (sample rMATS file and human chr 15 genome files)

$ pip install tox 
$ tox

To run JCAST using the test files and print the results to Desktop

$ python -m jcast {j}/tests/data/rmats {j}/tests/data/genome/Homo_sapiens.GRCh38.89.chromosome.15.gtf  {j}/tests/data/genome/Homo_sapiens.GRCh38.dna.chromosome.15.fa.gz -o ~/Desktop

where {j} is replaced by the path to JCAST.

Example Usage

The following is an example using JCAST to generate custom databases from ENCODE public RNA-seq dataset to generate a cardiac-specific database with JCAST.

Download RNA-Seq from ENCODE:

As an example, we will download the .fastq files from ENCODE adult human heart dataset 1 and dataset 2.

Align the FASTQ files to a reference genome

Read alignment can be performed using STAR >= v.2.5.0, e.g.,:

$ STAR --runThreadN 10 --genomeDir path/to/GRCh38/STARindex --sjdbGTFfile path/to/Homo_sapiens.gtf --sjdbOverhang 100 --readFilesIn ./ENCFF781VGS.fastq.gz ./ENCFF466ZAS.fastq.gz --readFilesCommand zcat --outSAMtype BAM SortedByCoordinate --outFileNamePrefix ./STAR_aligned/b1t1/
$ STAR --runThreadN 10 --genomeDir path/to/GRCh38/STARindex --sjdbGTFfile path/to/Homo_sapiens.gtf --sjdbOverhang 100 --readFilesIn ./ENCFF731CDK.fastq.gz ./ENCFF429YOS.fastq.gz --readFilesCommand zcat --outSAMtype BAM SortedByCoordinate --outFileNamePrefix ./STAR_aligned/b2t1/

Note: Arguments including runThreadN and sjdbOverhang should be customized to suit your system and data files. Please refer to the STAR documentations for details.

Identify transcript splice junctions

Splice junctions can be found using rMATS with the .bam files following STAR. Please refer to the rMATS instructions for latest commands. The following example was tested using rmats-turbo-0.1 running in Docker and using rMATS v.4.1.0/Python 3.7. Support for stringtie assembled transcripts will be implemented in a future version.

Set up a Virtual Environment for rMATS turbo 0.1 in Python 2.7 (only if needed)

Install the rMATS image

Follow instructions from rMATS and docker specific to your OS. E.g.:

$ sudo docker load -i rmats-turbo-0.1.tar

Prepare the /rMATS subdirectory

Copy the individual .bam files from STAR into the rMATS subdirectory and rename them b1t1.bam, b1t2.bam, b2t1.bam, b2t2.bam, etc. Copy the GTF file from the Genomes folder as GRCm38.gtf. Write a b1.txt file with a text editor containing the following docker virtual directories:

/data/b1t1.bam,/data/b1t2.bam

Write a b2.txt file

/data/b2t1.bam,/data/b2t2.bam

Go back to the data directory and run the rMATS image. The -v flag mounts the host directory into the docker container at /data, which corresponds to the visual directories in the b1.txt and b2.txt files.

$ sudo docker run -v path/to/data/directory:/data rmats:turbo01 --b1 /data/b1.txt --b2 /data/b2.txt --gtf /data/GRCh38.gtf --od /data/output -t paired  --nthread 4 --readLength 101 --anchorLength 1

Note

Note: Arguments including nThread, readLength, and anchorLength should be customized to suit your system and data files. Please refer to the rMATS documentations for details.

Run the JCAST Python program specifying the path to the rMATS output directory, the genome sequence, as well as the GTF annotation file:

$ python -m jcast path/to/rMATS/output/encode_human_heart/ path/to/gtf/Homo_sapiens.GRCh38.89.gtf path/to/genome/Homo_sapiens.GRCh38.dna.primary_assembly.fa -o encode_human_heart