LABRAT

Method workflow for LABRAT

LABRAT is for quantifying the alternative usage of polyadenylation and cleavage sites in RNAseq data. You can find the publication. source code: https://github.com/TaliaferroLab/LABRAT

The quantification step of LABRAT does not qualify for the benchmark quantification, which requires TPM values for each polyA site. LABRAT only provides TPM per transcript. However, the output could be generated in the future by collecting the polyA site coordinates.

Input & pre-processing

• samplesheet.csv with the following columns (more_samplesheets/samplesheet_paired.csv was used for testing locally)
  • sample -- unique names of the samples
  • fastq1 -- single-end fastq or paired-end fastq1
  • fastq2 -- paired-end fastq2 (if single-end, please leave it empty)
  • gff -- gencode gff file (tested Ensembl does not work due to the lack of the gene_type field)
  • fasta -- genome fasta file
  • condition -- condition of the samples (requires at least two conditions)

Params

• --input samplesheet.csv
• --run_quantification for running quantification
• --run_differential for running differential analysis
• --labrat_quant_dir location of the quantification results directory. Please provide this if --run_differential but not --run_quantification
• --conditionA the first condition in the condition column of the samplesheet.csv. Please provide this if --run_differential
• --conditionB the second condition in the condition column of the samplesheet.csv. Please provide this if --run_differential

The Nextflow workflow contains two profiles: docker and singularity. Either can be used for exeuction and they are defined in nextflow.config.

example commands

running quantification only

nextflow main.nf -profile docker --input <samplesheet.csv> --run_quantification

running differential analysis only

nextflow main.nf -profile docker --input <samplesheet.csv> --run_differential --conditionA <first condition> --conditionB <second condition> --labrat_quant_dir <full_path to labrat_quantification_results_dir>

running both quantification and differential analysis

nextflow main.nf -profile docker --input <samplesheet.csv> --run_quantification --run_differential --conditionA <first_condition> --conditionB <second_condition>

Run test data

The workflow does not contain the setup to run the test data in tests/test_data as the genome sequence is necessary. The following steps are meant to setup the test data in order to be comparable with other method workflows.

• Create siControl_R1_2genes.fastq.gz and siSrsf3_R1_2genes.fastq.gz from corresponding BAM files with samtools bam2fq as described here: tests/test_data/README.md.
• Use gencode_2genes_Chr_prefix.vM18.annotation.gff3 as genome annotation (gff).
• Download genome sequence of mm10 from Gencode, preferentially only chr6 and chr16 as only those are used in the test.
• Provide the files with absolute paths in a sample sheet, by e.g. adjusting more_samplesheets/samplesheet_single.csv.
  • Leave column fastq2 empty
  • sample: siControl_R1 and siSrsf3_R1
  • condition: siControl_ and siSrsf3
• Run the quantification and differential analysis with (use actual sample sheet):

nextflow main.nf --profile docker --input more_samplesheets/samplesheet_single.csv \
   --run_quantification --run_differential \
   --conditionA siControl --conditionB siSrsf3

Output & post-processing

LABRAT consists of the following three steps:

• makeTFfasta -- makes a fasta file of transcripts that will later be quantified by salmon
• runSalmon -- quantifies transcript abudance with salmon (please see Notes below)
  • each sample will have its own output directory
  • quant.sf is used for generating the quantification bed file but without chromStart and chromEnd inforamtion (the starting and ending position of the feature in the chromosome.)
• calculatepsi -- calculates the 𝜓 values for each gene in each sample and the identify genes that show significantly different 𝜓 values across conditions.
  • The output LABRAT.psis.pval is used for generation the differential usage tsv file

Notes

The LABRAT.py --mode runSalmon step is replaced by running salmon directly with the salmon index and salmon quant commands indicated in LABRAT.py due to c++ compling problems of salmon in LABRAT's BioContainer.