Segmentation fault during STAR mapping on specific fastq files (2.7.9a)

[angleYuan]

Hi All,

I would like to first thank you for introducing the multi-mapping features in STARsolo! I am exploring this feature now and it does make a significant difference to my data.

But I have a problem with one sample. I always encountered segmentation faults during the mapping step with this sample, no matter I ran it on a cluster or on a local server, playing with some parameters, etc. This sample has been sequenced on four lanes. And later I noticed that the problem lies with the fastq files from lane 4. If only the fastq files from lane 1,2,3 were provided, STARsolo finished without an error. On the contrary, running the STARsolo on lane 4 alone gave segmentation fault as well.

However, I cannot tell what the problem is with the fastq files from lane 4. I checked them with FastQC and run them with CellRanger v5, both finished without any problems.

I attached the STARsolo log.out and log.out.progess files from running on data from all 4 lanes (failed) and from only 3 lanes (success). I can provide other files if you think they will be helpful. I am looking forward to your suggestions. Thanks a lot!

Best,
Xuefei

3lanes.Log.final.out.txt
3lanes.Log.out.txt
3lanes.Log.progress.out.txt
4lanes.Log.out.txt
4lanes.Log.progress.out.txt

[alexdobin]

Hi Xuefei,

this looks like a bug. Would it be possible for you to share the FASTQ file from the 4th lanes (that seg-faulted)?

Thanks!
Alex

[angleYuan]

Hi Alex,

Thanks a lot for your help! You can download the 4th lane fastq files here (https://www.dropbox.com/sh/4u7vbh81ktdf5ry/AABj0meVj5J6dGMOSiYzVsc9a?dl=0). Please let me know if there is any problem.

Best regards,
Xuefei

[alexdobin]

Hi Xuefei,

thanks for the files! I have run them on genome generated from ENSEMBL FASTA and GTF, and could not reproduce the error. Could you please share the FASTA and GTF you have been using?

Thanks!
Alex

[angleYuan]

Hi Alex,

Thank you for your trial!

Yes we indeed used a reannotated gtf file (which you can download from the same dropbox folder https://www.dropbox.com/s/cduj7ss7rs1lnss/combined_annotation.gtf?dl=0). The fasta file was just downloaded from Ensembl (http://ftp.ensembl.org/pub/release-103/fasta/oryctolagus_cuniculus/dna/Oryctolagus_cuniculus.OryCun2.0.dna_sm.toplevel.fa.gz). The reference that we made from them worked for cellranger but not STARsolo with the 4th lane data.

Thanks a lot for looking into this.

Best,
Xuefei

[alexdobin]

Hi Xuefei,

I think I have fixed this problem. Would you mind trying to run your samples with the latest commit in the https://github.com/alexdobin/STAR/tree/dev_ExonOverIntron
branch.
I will make a 2.7.10a release in a few days.

Cheers
Alex

[angleYuan]

Thanks a lot for your help! Yes the new version works perfectly for me.