PacBio iso-seq data analysis¶

Summary¶

Pipeline adopted from ENCODE Long Read RNA-Seq Analysis Protocol for Human Samples (v3.0): https://www.encodeproject.org/documents/81af563b-5134-4f78-9bc4-41cb42cc6a48/@@download/attachment/ENCODE%20Long%20Read%20RNA-Seq%20Analysis%20Pipeline%20v3%20%28Human%29.pdf

Only work for hg38 right now, by 6/6/2022.

Current pipeline doesn’t filter/merge based on replicates.

Iso-seq read structure¶

Here at St. Jude we use a double barcoded primers, each sample has the unique barcode, 5p and 3p primers have the identical but reverse completed barcodes.

Input¶

input.list¶

A 2-column tsv file, the first column is PacBio hifi reads, the last column is sample ID, should be unique.

control.bam     control
KO.bam  KO

primer.fa¶

>primer_5p
GCAATGAAGTCGCAGGGTTGGGG
>primer_3p
GTACTCTGCGTTGATACCACTGCTT

Usage¶

hpcf_interactive

module load python/2.7.13

run_lsf.py -f input.list --primer primer.fa -p iso_seq

Output¶

Results can be found in the $jid folder.

1. raw read annotation file¶

*final_talon_read_annot.tsv

2. transcript/gene abundance file (main results)¶

*final_talon_abundance.tsv

3. custom gene annotation file¶

*talon.gtf

Reference¶

https://github.com/PacificBiosciences/IsoSeq

Notes¶

To resume failed runs, submit the job with the same jid -j and also use --override_jid

code @ github.