Skip to content

gxiaolab/isoLASER

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

33 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

isoLASER

github

Hi, Thank you for your interest in our tool. Please be advised that the corresponding manuscript is under review, and we will be updating the source code based on the comments from the reviewers.

About

isoLASER performs gene-level variant calls, phasing, and splicing linkage analysis using third-generation RNA sequencing data.

Table of contents

Requirements

isoLASER is written in Python 3.8 and requires the following packages (with the tested versions):

  • multiprocess
  • numpy==1.24.4
  • scipy==1.10.1
  • pandas==2.0.3
  • pysam==0.16.0.1
  • pytabix==0.1
  • pyfaidx==0.5.8
  • biopython==1.76
  • vcfpy==1.0.3
  • HTSeq==0.12.4
  • networkx==3.1
  • scikit-learn==1.3.2

isoLASER has been tested in the following operating systems:

  • CentOS Linux 7

External software requirements:

Installation

You can clone this GitHub repository:

git clone [email protected]:gxiaolab/isoLASER.git 
cd isoLASER
python -m build
pip install .

It is recommended to install isoLASER in a virtual environment.

conda create -n isoLASER_env python=3.8

You can also download the Singularity container:

singularity pull library://giovas/collection/s6
singularity exec s5_latest.sif isoLASER

If successful, the program is ready to use. The installation incorporates console script entry points to directly call isoLASER:

isoLASER --help

Installation time varies depending on the number of dependencies that need to be installed. Assuming all library dependencies are installed already, the installation of isoLASER should only take a few seconds.

Preprocessing

Long-read RNA sequencing is notorious for its high base-calling error rate. As such, it is important to clean and preprocess the data to discard false transcripts resulting from misalignment, bad consensus, truncation, and other technical artifacts.

Transcript identification

isoLASER needs a GTF file as input to annotate individual reads with their isoform membership. Ideally, this GTF file is built using a long-read annotation software such as Talon, Clair, Bambu, Espresso, or similar.

Details of the Talon pipeline can be found on their GitHub repository.TALON.

  • The pipeline consists of correcting alignments around splice junctions using TranscriptClean, and labeling the reads for internal priming using talon_label_reads.
  • The processed bam files are then ready to be annotated by first creating a database with talon_initialize_database and then annotating the individual reads with talon.
  • Finally, the transcripts are filtered with talon_filter_transcripts and a GTF file is constructed with the retained transcripts with the command talon_create_GTF.

Annotate bam file

From the annotation used in the previous step, use the GTF file to generate a transcriptome reference for alignment.
This step serves to assign transcript ids to every read of the target bam file.

# generate transcriptome reference from GTF
isolaser_convert_gtf_to_fasta -g {annot.gtf} -f {reference.fa} -o {transcriptome}

# convert bam file to fastq for re-alignment
samtools fastq {input.bam} > {input.fq}

# re-align against newly generated transcriptome reference
minimap2 -t 16 -ax splice:hq -uf --MD {transcriptome.fa} {input.fq} > {transcriptome.sam}

The output is a sam file where the contigs are transcript ids. Next, filter for secondary, supplementary and trans-gene reads whilst annotating with transcript ids.

isolaser_filter_and_annotate -b {input.bam} -t {transcriptome.sam} -g {annot.gtf} -o {input.annot}

The output is a bam file: input.annot.bam after some basic filtering (secondary and supplemental reads), trans-gene filtering and the ZG and ZT tags with the name of the corresponding gene and transcript id for each read.

Extract exonic parts from GTF

isoLASER uses an exon-centric approach to analyze splicing and exonic-parts are a great granular approach to understand local splicing changes.

isolaser_extract_exonic_parts -g {annot.gtf} -o {transcript.db}

The output is a new directory transcript.db that contains a pickle file per gene encapsulating all the exonic parts and transcripts associated with them.

Run isoLASER

isoLASER requires the annotated bam file (with ZG and ZT tags), the transcriptome database with exonic parts, and a reference annotation (e.g. hg38.fa).

isoLASER -b {input.annot.bam} -o {output.prefix} -t {transcript.db} -f {reference.fa}

# output:
{output.prefix}.vcf
{output.prefix}.mi_summary.tsv

The output is very extensive and includes information that is only relevant for the joint analysis or plotting. To obtain the significant allele-specific events (cis-directed splicing events) use the filter function:

isolaser_filter -m {output.prefix.mi_summary.tsv} -o {output.prefix.mi_summary.filtered.tsv}

Run isoLASER joint

The first step is a wrapper of GATK functions to merge the variant calls from different samples.

The input file fofn contains informaiton of the individual samples

# fofn.tsv
SM1 SM1.bam SM1.mi_summary.tsv
SM2 SM2.bam SM2.mi_summary.tsv

Run the combine vcf step:

isoLASER_combine_vcf -f {fofn.tsv} -o {output.prefix}

# output:
{output.prefix}.combined.vcf
{output.prefix}.genotyped.vcf

Perform joint analysis

isoLASER_joint -f {fofn.tsv} -o {output.prefix} -t {transcript.db}

# output:
{output.prefix}.merged.mi_summary.tsv

Make a nigiri plot

Parse the .mi_summary.tsv file to obtain the list of events to plot

nigiri_parse --mi {output.prefix.mi_summary.tsv} -o {output.plot} -t {transcriptome.db} 

# output:
{output.plot}.cis_events.bed
{output.plot}.cis_genes.tsv

Split the bam file

nigiri_split -b {input.annot.bam} -v {var.string} -o {fofn} 

We use an adapted version of ggsashimi

nigiri_plot -b {fofn} -c {region} -o {output.plot} 

FAM221A

Demo

For a complete demo please check the test_pipeline repository.

Debug

If you experience any issues please submit your question to the Issues tab on this website.