cufflinks- cuffquant issue

Question

[soyeon]

root@86-empty:/data/soyeon/tool/cufflinks-2.2.1.Linux_x86_64# ./cuffquant --no-update-check -o /data/soyeon/20240315_RNA_FPKM/cxb -p 10 --library-type fr-firststrand /data/soyeon/reference_genome/hg38.refGene.gtf /data/soyeon/20240315_RNA_FPKM/mapping_bam/2+Aligned.sortedByCoord.out.bam
[11:31:41] Loading reference annotation.
[11:31:51] Inspecting maps and determining fragment length distributions.

The issue arises during the "Inspecting maps and determining fragment length distributions." stage of the "cuffquant" command, where the process takes a considerable amount of time. It takes more than an hour. It originally took about 7 minutes per sample, but it is strange that it now takes over an hour.

Does it take this long? Somebody help me, please...

[dsull]

Why do people still use cuff-whatever? Just curious.

[Rob]

IME, mostly because that's what their lab's pipeline is. Often, the PI doesn't understand the critical importance of upgrading these computational pipelines, though they no doubt understand the benefits of using improved biotechnologies and reads greater than 35bp. There is a very serious asymmetry in propagating the understanding of improved computational methods versus other types of technological or methodological improvements.

Answer 1

[stern_am_himmel]

This happens sometimes. I'm not 100% sure, but it seems to occur when the program encounters a gene that has a high number of 'bundle frags.' Try reducing the max-bundle-frags parameter and check the 'HIDATA' gene. That particular gene slows down the program. If it isn't important, maybe you can remove it from the GTF file