genomeGenerate failure after packing SA

Gabe Zentner

unread,

Apr 6, 2016, 1:08:18 PM4/6/16

to rna-star

Hi there,

I would like to use STAR to map reads to the mm10 genome assembly, but I have had a recurrent problem building a STAR index. Using 16 cores and 32 GB of RAM, STAR makes it here:

Apr 06 10:42:35 ..... Started STAR run

Apr 06 10:42:35 ... Starting to generate Genome files

Apr 06 10:43:39 ... starting to sort Suffix Array. This may take a long time...

Apr 06 10:43:57 ... sorting Suffix Array chunks and saving them to disk...

Apr 06 10:51:41 ... loading chunks from disk, packing SA...

It stops with SA_49 left in the directory. It appears that the job was killed, based on the other output file:

star version 2.4.1d loaded.

/var/spool/torque/mom_priv/jobs/1379849.m2.SC: line 7: 9831 Killed STAR --runThreadN 16 --runMode genomeGenerate --genomeDir /N/dc2/scratch/gzentner/mm10_star --genomeFastaFiles /N/u/gzentner/Karst/genomes/mm10_files/mm10_cat.fa --sjdbGTFfile /N/u/gzentner/Karst/genomes/mm10_files/mm10_refGene.gtf

The log file is attached. Thanks in advance for any advice!

Gabe

Log.out

Alexander Dobin

unread,

Apr 6, 2016, 6:18:24 PM4/6/16

to rna-star

Hi Gabe,

this is most likely due to some shortage in RAM. Are all of 32GB available on your server? Do you have access to a server with larger RAM, my guess you would need just 2-3 GB more?

Also, some users reported RAM problems with pre-compiled STAR executables, could you try to compile STAR with make?

Cheers

Alex

Gabe Zentner

unread,

Apr 9, 2016, 11:36:01 PM4/9/16

to rna-star

Hi Alex,

I was able to rerun it on another of our servers with 40 GB of RAM, no problem. Thanks for the advice!

Gabe

Reply all

Reply to author

Forward