Problem with talon_initialize_database

[Kiliankleemann]

Tried to run talon_initialize_database but got an error:

talon_initialize_database --f  reference/GRCh38_GENCODE_rmsk_TE_reformatted.gtf \
  --g hg38_rmsk_ucsd \
  --a hg38 \
  --o hg38 
chrY
bulk update genes...
bulk update gene_annotations...
Traceback (most recent call last):
  File "/home/kilian/anaconda3/envs/talon/bin/talon_initialize_database", line 8, in <module>
    sys.exit(main())
  File "/home/kilian/anaconda3/envs/talon/lib/python3.7/site-packages/talon/initialize_talon_database.py", line 1073, in main
    populate_db(db_name, annot_name, chrom_genes, chrom_transcripts, exons, genome_build)
  File "/home/kilian/anaconda3/envs/talon/lib/python3.7/site-packages/talon/initialize_talon_database.py", line 634, in populate_db
    add_transcripts(c, transcripts, annot_name, gene_id_map, genome_build)
  File "/home/kilian/anaconda3/envs/talon/lib/python3.7/site-packages/talon/initialize_talon_database.py", line 743, in add_transcripts
    db_gene_id = gene_id_map[native_gene_id]
KeyError: 'AluY'

[Kiliankleemann]

I made sure the reformatting of GTF is correct:

wget https://hgdownload.soe.ucsc.edu/goldenPath/hg38/database/rmsk.txt.gz
gzip -d *.gz
talon_reformat_gtf -g reference/GRCh38_GENCODE_rmsk_TE.gtf

talon_initialize_database --f reference/GRCh38_GENCODE_rmsk_TE_reformatted.gtf \
  --g hg38_rmsk_ucsd \
  --a hg38 \
  --o hg38 

[fairliereese]

Would you be able to share the GTF that you're using with me? I will try running it on my end and see if I can pinpoint the issue.

[Kiliankleemann]

Should be able to download the gtf and unzp with the first command - thats the one I tried

[fairliereese]

This one? https://hgdownload.soe.ucsc.edu/goldenPath/hg38/database/rmsk.txt.gz This does not look like a GTF to me. For example, the strand should be in the 6th column (0-indexed), but looks like it's in the 9th column of your file.

[Kiliankleemann]

Which gtf did you use for hg38 repeatmasker?

[fairliereese]

For me to best help you, you should send all the commands that you used to download / format your GTF. I think I'm missing some information from your side.

[sojichld]

I'm having a similar issue and I'm not really sure why. I've also tried using the gtf formatter with no luck.

It took 0:00:00.01 to process chromosome
NW_023397527.1
Traceback (most recent call last):
File "/users/aademilu/.local/bin/talon_initialize_database", line 8, in
sys.exit(main())
File "/users/aademilu/.local/lib/python3.8/site-packages/talon/initialize_talon_database.py", line 1015, in main
populate_db(db_name, annot_name, chrom_genes, chrom_transcripts, exons, genome_build)
File "/users/aademilu/.local/lib/python3.8/site-packages/talon/initialize_talon_database.py", line 596, in populate_db
transcripts = chrom_transcripts[chromosome]
KeyError: 'NW_023397527.1'

I've attached an example of the file. The full file can be found here.
gtf_example.txt

[fairliereese]

Can you please send me the exact command you tried for talon_initialize_database, as well as the version number of TALON that you're using?

[sojichld]

Can you please send me the exact command you tried for talon_initialize_database, as well as the version number of TALON that you're using?

talon_initialize_database --f ../../reference/GCF_004126475.2_mPhyDis1.pri.v3_genomic.gtf --a discolor_annot --g discolor --o discolor

Where can I find version information?

[fairliereese]

I don't think there's a nice way to access the version info now, but if you haven't updated TALON in a long time it might be worth pulling and installing the latest commits. On my machine, I am able to run your init command with gtf_example.txt no problem. Did you also verify that you're having an issue with the small file too?

[sojichld]

Yes, while that one does run for me as well (it doesn't inlcude NW_023397527.1 ), I cannot get other cuts of the file to work, it creates an error as follows:

    genes, transcripts, exons = read_gtf_file(gtf_file)
  File "/users/aademilu/.local/lib/python3.8/site-packages/talon/initialize_talon_database.py", line 495, in read_gtf_file
    entry_type = tab_fields[2]

I noticed that the gene is the only one of that scaffold, maybe that could be the issue? I have provided the full file, which will run until the scaffold in question. The program will run if I remove the gene from the gtf.

[fairliereese]

The problem is that gene does not have any transcripts annotated to it. If you look, it goes from one gene entry (the one on your NW_023397527.1 chromosome) to the next, without any additional entry. I would advise removing this entry from your GTF and moving on with your analysis.

[sojichld]

You're right. Strange. I've removed it an it works just fine. Was able to fully process everything. Thanks.