WARNING - Error, incorrect gff. exon doesn't have valid transcript and Intron offset 6376 miscalculated

[Xing]

Hello,

When I run the majiq, I met the following problem:

command :

majiq build  -c ./config_new.txt /data/mus/ref/ensembl/GFF3/Mus_musculus.GRCm39.104.gff3 -o ./build/  -j 30

config file:

[info]
bamdirs=/data/mus/nanopore/top3b_brain_majiq/bams
sjdirs=./sj
genome=mm10
strandness=none
[experiments]
WT=1.aln.sorted,2.aln.sorted,3.aln.sorted,4.aln.sorted
KO=5.aln.sorted,6.aln.sorted,7.aln.sorted,8.aln.sorted

Errors:

2024-07-06 17:03:44,041 (PID:1767952) - WARNING - Error, incorrect gff. exon doesn't have valid transcript b'transcript:ENSMUST00000166255'
2024-07-06 17:03:44,050 (PID:1767952) - WARNING - Error, incorrect gff. exon doesn't have valid transcript b'transcript:ENSMUST00000095364'
2024-07-06 17:03:44,193 (PID:1767952) - WARNING - Error, incorrect gff. exon doesn't have valid transcript b'transcript:ENSMUST00000103555'

.....

Intron offset 6376 miscalculated
Intron offset 12775 miscalculated
Intron offset 4429 miscalculated
Intron offset 5748 miscalculated
Intron offset 3781 miscalculated
Intron offset 3823 miscalculated
Intron offset 5180 miscalculated
.........

Any suggestions?

thank you.

[San Jewell]

Hello,

In general this indicates that something strange might have occurred in the specific experiment which yields unexpected or undefined results. I don't have a good suggestion for clearing that warning without seeing the data to determine what might have happened. Also note that this is a warning and not an error, so if the problem only affects a few genes you can use the majority of the results without an issue.

Do you think there is and chance we might be able to chat in private to check on this specific data to see what is happening?

Thank you,

-San

[Xing]

Yes, sure. Thank you. Nice to talk with you. How about July 24 10:00 am? Or you can give a time slot, I will check whether it's available for me.

Xing

--
You received this message because you are subscribed to a topic in the Google Groups "Biociphers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/majiq_voila/HeRHfNYAGY8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to majiq_voila...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/majiq_voila/f014bc31-fa0d-4ef5-b162-0b14f7a5ea97n%40googlegroups.com.

[Xing]

Hi, 

I want to say that it's not several genes but lots of genes, even hundreds to thousands. That's why I post it here and still wait for further feedback to do the following majiq analysis.

Xingliang

[San Jewell]

Hi Xingliang, I'm sorry I was away at a conference and I didn't see your initial suggestion fast enough.

Would you or folks you work with be opposed to transferring a subset of the data to us for review, so that we can reproduce and debug the specific cause of this issue? If so I can communicate privately with you to arrange it.

Thanks,

-San