Hi all,
I have generated the genome using STAR and it works fine. Now I try to map PE reads (size 100000 reads for each file) to the generated genome using the following command:
srun STAR --genomeDir $hg_ref_path --outFileNamePrefix $output_prefix \
--seedSearchStartLmax 30 --outFilterMultimapNmax 10 --outFilterMismatchNmax 12 --alignMatesGapMax 1000 --outSAMattrIHstart 0 \
--outSAMmapqUnique 255 --outFilterIntronMotifs RemoveNoncanonical --outFilterScoreMinOverLread 0.4 --outFilterMatchNminOverLread 0.4 \
--chimSegmentMin 20 --outSAMtype BAM Unsorted --runThreadN 4 --readFilesIn $file_R1 $file_R2
but I still get the segmentation fault error:
May 09 00:20:27 ..... Started STAR run
May 09 00:20:27 ..... Loading genome
May 09 00:21:37 ..... Started mapping
srun: error: compute-1-14: task 0: Segmentation fault
I tried also to include the GTF annotation file with the --sjdbGTFfile option but the same error was occured:
May 09 01:04:01 ..... Started STAR run
May 09 01:04:01 ..... Loading genome
May 09 01:05:10 ..... Processing annotations GTF
srun: error: compute-1-14: task 0: Segmentation fault
Any suggestions about these errors ?
How much RAM do you have and/or you are allocating to this job? STAR requires a significant amount of memory (~30+G for human sized genomes).
I allocated 45G of RAM to this job