STARsolo seg fault without readMapNumber flag

Question

[rbronste]

Hi,

So I'm trying to use STARsolo on a couple of datasets and having strange issue, when running the following command via slurm and NOT using the --readMapNumber flag I always get a segmentation fault when the counting begins irrespective of different thread settings and cluster memory settings etc:

`#!/bin/bash`
#SBATCH --job-name=STAR_index
#SBATCH --output=/gpfs/projects/STARindex/
#SBATCH --ntasks-per-node=1
#SBATCH --nodes=4
#SBATCH --time=48:00:00
#SBATCH --partition long-28core
#SBATCH -o STAR_index.out
#SBATCH --mail-type=BEGIN,END`

STAR --runThreadN 20 --genomeDir /gpfs/projects/STAR --readFilesIn /gpfs/projects/RNA_data/SC292_RNAfastq/SC292RNA_S6_L003_R2_001.fastq.gz /gpfs/projects/RNA_data/SC292_RNAfastq/SC292RNA_S6_L003_R1_001.fastq.gz --runDirPerm All_RWX --soloType CB_UMI_Simple --soloBarcodeReadLength 150 --soloCBwhitelist /gpfs/projects/cellranger_arc/cellranger-arc-2.0.0/lib/python/cellranger/barcodes/737K-arc-v1.txt --soloCBmatchWLtype 1MM_multi_Nbase_pseudocounts  --soloUMIdedup 1MM_CR --soloUMIfiltering MultiGeneUMI_CR --soloCellFilter EmptyDrops_CR --clipAdapterType CellRanger4 --outFilterScoreMin 20 --outSAMtype BAM SortedByCoordinate --outSAMattributes CR UR CY UY CB UB --soloFeatures Gene GeneFull SJ Velocyto --soloOutFileNames output/ features.tsv barcodes.tsv matrix.mtx --soloMultiMappers EM --readFilesCommand zcat --readMapNumber 20938859 --outReadsUnmapped Fastx

However if I do the following for each fastq file:

echo $(cat /gpfs/projects//RNA_data/SC292_RNAfastq/SC292RNA_S6_L003_R1_001.fastq.gz|wc -l)/4 | bc

and add the total reads for both fastq files via --readMapNumber flag it runs through without issues, so just wondering if this approach is ok and why it might be seg faulting without that flag. Thanks.

[Ram]

Long shot but try this - your .gz files don't look like they're gzipped. Either gzip them or rename them to not have the .gz suffix then try the STAR command without the --readMapNumber parameter.

[rbronste]

Have tried it both ways - gzipped or not and still seg-faults unfortunately. Wondering if its correct in this instance to add the total number of reads across the two fastq files for the --readMapNumber flag? Its weird because even if I put in 100M reads for that flag, which is 5 times greater than what was done it does not seg fault, only does without flag at all. Thanks.

[Ram]

What version of STAR are you using? You might be better off opening a GitHub issue on alexdobin/STAR.

[rbronste]

Im using STAR version: 2.7.9a

Thanks already posted as well on alexdobin/STAR

[GenoMax]

Your files have identical number of reads and they are in sync. Just checking.

[rbronste]

Just checking? What do you mean?

[GenoMax]

Checking to see if you have verified that your files have an equal number of reads in R1/R2 files and they are in order in both files.

[rbronste]

Oh ok got it, misiunderstood what you meant. Yes to both of those.

[Ram]

Please edit your post and link to the GitHub issue. Also edit the GitHub issue and add a link to this post.