CuffDiff hangs at same point

Question

[laboiss]

Hi all,

I am currently encountering errors with CuffDiff because it appears to be stuck at the same locus for the past 48 hours and I was wondering if anyone else had this problem or had found a solution?

Here is my cuffdiff command line: cuffdiff -o diff_out -b index_mouse.fa -p 8 -L M1,M2 merged_asm/merged.gtf ./M1_1_3.th/accepted_hits.bam,./M1_2_13.th/accepted_hits.bam,./M1_2_14.th/accepted_hits.bam,./M1_2_15.th/accepted_hits.bam,./M1_2_17.th/accepted_hits.bam ./M2_1_2.th/accepted_hits.bam,./M2_2_8.th/accepted_hits.bam,./M2_2_23.th/accepted_hits.bam,./M2_2_24.th/accepted_hits.bam,./M2_2_26.th/accepted_hits.bam

The point where this command is stuck is under "Testing for differential expression and regulation in locus".

Thanks!

Lauren

[Fabio Marroni]

Does "Stuck at the same locus" means that it is analysing one single gene and takes forever? Keep in mind that cuffdiff can take long to run. What is the CPU usage? However, the best thing you can do is to subset the problem (IMHO this is ALWAYS the best thing) i.e. feed to the software a subset of the bams, i.e. only a couple of them, or even better create very small bam in which you only have regions close to the problematic locus and see if it happens again.