We have a very specific problem with cufflinks v.2.1.1 on a Rocks cluster (v.5.3). We have been getting segmentation faults with cufflinks v.2.1.* releases.
The error message is not very informative (even when run with -v option). Following is an excerpt from the test data run included in the Trapnell Nature Protocols paper (this example works fine with Cufflinks v. 2.0.2)
Our system administrators have rebuilt the relevant cufflinks module (we use modules to manage software) multiple times with various Boost library/Eigen combinations. We are stumped at this point as to what to try next.
Is there someone on this forum who has cufflinks v.2.1.1 working on Rocks (v.5.3)?
The error message is not very informative (even when run with -v option). Following is an excerpt from the test data run included in the Trapnell Nature Protocols paper (this example works fine with Cufflinks v. 2.0.2)
Code:
2L:7534-9404 Assessing overlaps between 1 fragments for identical conflict sets 2L:7534-9404 Compressed 0 of 1 redundant fragments (0.000000 percent) 2L:7534-9404 Ignoring 25 strictly contained fragments 2L:7534-9404 Assembling bundle with 1 hits 2L:7534-9404 Calculating scaffold densities 2L:7534-9404 Creating compatibility graph 2L:7534-9404 Performing final collapse round Extracted 1 contiguous transfrags from 1 scaffolds 2L:7534-9404 Starting new collapse round line 8: 7316 Segmentation fault cufflinks -p 4 --no-update-check -v -o GSM794483_C1_R1_clout_new GSM794483_C1_R1_thout/accepted_hits.bam
Is there someone on this forum who has cufflinks v.2.1.1 working on Rocks (v.5.3)?
Comment