Getting frequent Segmentation fault errors

[sanyalab]

Hello,

I am trying to run the following command and getting "Segmentation Fault" errors. Below is the error

sh: line 1: 26387 Segmentation fault (core dumped) /scale03/fs0/gpfs0/research/anno/sanyalab/SOFTWARE/psiclass/combine-subexons --ls ./subexon/Hap1_psiclass_subexon.list > ./subexon/Hap1_psiclass_subexon_combined.out Terminated

My Command was nohup /anno/sanyalab/SOFTWARE/psiclass/psiclass -b STAR-HAP1/egg_lateAligned.sortedByCoord.out.bam,STAR-HAP1/Instar_1_Aligned.sortedByCoord.out.bam,STAR-HAP1/Instar_2_Aligned.sortedByCoord.out.bam,STAR-HAP1/Instar_3_Aligned.sortedByCoord.out.bam,STAR-HAP1/Instar_3_gut_Aligned.sortedByCoord.out.bam,STAR-HAP1/pupa_Aligned.sortedByCoord.out.bam,STAR-HAP1/unmated_adult_female_Aligned.sortedByCoord.out.bam,STAR-HAP1/unmated_adult_male_Aligned.sortedByCoord.out.bam -p 14 --vd 1.5 --stage 0 -o Hap1_psiclass

Each BAM file is between 9GB - 13GB. I have a RAM of 256GB. I have done alignment using STAR. This command worked fine with HISAT2. HISAT2 alignment files were also between 9GB - 13GB.

I can use the cluster and specify a higher memory. However, I do not know whether I should be giving the complete path to the prefix in "-o" like so "-o /Full/Path/2/Out_prefix"

Please advice what I should be doing. I am on an ubuntu x86_64 machine.

Thanks Abhijit

[mourisl]

Which version of PsiCLASS are you using? Could you please share the subexon files in the subexon folder for me to debug with? Thanks.

[sanyalab]

Hello,

I am using the latest psiclass v1.0.2. The subexon folder is 218 MB. Let me know if you are able to get the same.

Thanks Abhijit subexon.zip https://drive.google.com/file/d/1pC5KR-DxfpN-RmgZDqx9JtAOQEvxSoo4/view?usp=drive_web

On Mon, Apr 18, 2022 at 8:53 AM Li Song @.***> wrote:

Which version of PsiCLASS are you using? Could you please share the subexon files in the subexon folder for me to debug with? Thanks.

— Reply to this email directly, view it on GitHub https://github.com/splicebox/PsiCLASS/issues/24#issuecomment-1101055688, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE545CMKWYMIGR6CUGB5RITVFTIR5ANCNFSM5TUY5TIA . You are receiving this because you authored the thread.Message ID: @.***>

[mourisl]

It seems the current github version does not crash on our server. Could you please use git pull or git clone to try the most updated version? If it resolves your issues, I will draft a new release. Thank you.

[sanyalab]

Ok! Doing this now. Will post on the results. Thank you so much.

-Abhijit

[sanyalab]

Getting Segmentation fault again. Seems to be working fine with HISAT2 created BAM files. I get the desired output files. Weird. Because if the STAR output files are bad then Stringtie, scallop will also not work.

-Abhijit

[mourisl]

Could you please recompile PsiCLASS with the debug mode : "make clean" and then "asan=1 make"? This time if the program crashes, it should print which part causes the issue. To save time, you can run psiclass with the option "--stage 2" to directly start from the combine subexon step. Thank you.

[sanyalab]

Hi, Here is what is happening.

lxjh218:.../SOFTWARE/psiclass$ make clean rm -f *.o *.gch subexon-info combine-subexons trust-splice vote-transcripts junc grader add-genename addXS

lxjh218:.../SOFTWARE/psiclass$ asan=1 make g++ -c -o subexon-info.o -I./samtools-0.1.19 -L./samtools-0.1.19 -Wall -O3 -fsanitize=address -g SubexonInfo.cpp -lbam -lz -lm -lpthread -fsanitize=address -ldl -g cc1plus: error: unrecognized command line option "-fsanitize=address" cc1plus: error: unrecognized command line option "-fsanitize=address" make: *** [subexon-info.o] Error 1

So then I did just make. Attached is the STDOUT. make_for_psiclass.txt

Thanks Abhijit

[mourisl]

What is the gcc version on your system? You may need to use a newer version.

[sanyalab]

lxjh218:.../SOFTWARE/psiclass$ gcc --version gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-23)

[sanyalab]

I installed psiclass using conda. This time the error is different. Below

sh: /Anaconda3/envs/SRAssembly1/bin/samtools-0.1.19/samtools: No such file or directory Found mate read id index suffix(.1 or /1). Calling "--mateIdx 1" option. If this is a false calling, please use "--mateIdx 0". /Anaconda3/envs/SRAssembly1/bin/junc /STAR-RESULT/Sample1_Aligned.sortedByCoord.out.bam -a --hasMateIdSuffix > ./splice/Hap1_psiclass_bam_0.raw_splice /Anaconda3/envs/SRAssembly1/bin/junc /STAR-RESULT/Sample2_Aligned.sortedByCoord.out.bam -a --hasMateIdSuffix > ./splice/Hap1_psiclass_bam_1.raw_splice /Anaconda3/envs/SRAssembly1/bin/junc /STAR-RESULT/Sample3_Aligned.sortedByCoord.out.bam -a --hasMateIdSuffix > ./splice/Hap1_psiclass_bam_2.raw_splice /Anaconda3/envs/SRAssembly1/bin/junc /STAR-RESULT/Sample4_Aligned.sortedByCoord.out.bam -a --hasMateIdSuffix > ./splice/Hap1_psiclass_bam_3.raw_splice /Anaconda3/envs/SRAssembly1/bin/junc /STAR-RESULT/Sample4_V1_Aligned.sortedByCoord.out.bam -a --hasMateIdSuffix > ./splice/Hap1_psiclass_bam_4.raw_splice /Anaconda3/envs/SRAssembly1/bin/junc /STAR-RESULT/Sample5_Aligned.sortedByCoord.out.bam -a --hasMateIdSuffix > ./splice/Hap1_psiclass_bam_5.raw_splice /Anaconda3/envs/SRAssembly1/bin/junc /STAR-RESULT/Sample6_Aligned.sortedByCoord.out.bam -a --hasMateIdSuffix > ./splice/Hap1_psiclass_bam_6.raw_splice /Anaconda3/envs/SRAssembly1/bin/junc /STAR-RESULT/Sample7_Aligned.sortedByCoord.out.bam -a --hasMateIdSuffix > ./splice/Hap1_psiclass_bam_7.raw_splice /Anaconda3/envs/SRAssembly1/bin/trust-splice ./splice/Hap1_psiclass_splice.list /STAR-RESULT/Sample1_Aligned.sortedByCoord.out.bam > ./splice/Hap1_psiclass_bam.trusted_splice perl /Anaconda3/envs/SRAssembly1/bin/FilterSplice.pl ./splice/Hap1_psiclass_bam_0.raw_splice ./splice/Hap1_psiclass_bam.trusted_splice > ./splice/Hap1_psiclass_bam_0.splice Can't open perl script "/Anaconda3/envs/SRAssembly1/bin/FilterSplice.pl": No such file or directory Terminated

I think I need samtools v0.1.19 to be installed in conda.

[sanyalab]

I installed samtools v0.1.19 in conda. It still is looking for samtools at the following address

/mnt/common/rh6/annotate/Anaconda3/envs/SRAssembly1/bin/samtools-0.1.19/samtools but samtools is present at /mnt/common/rh6/annotate/Anaconda3/envs/SRAssembly1/bin and it is the first entry in PATH.

Samtools needs to be packaged correctly in conda I think.

-Abhijit

[mourisl]

Can you use the command like "export PYTHONUSERBASE=intentionally-disabled" to force the system in conda?