illegal character 0

[dptech-corp]

illegal character 0

[dominik-handler]

I was able to fix my problem by manually removing all sequences containing invalid characters according to this:
google-deepmind/alphafold#569

tac pdb_seqres.txt | sed '/^CT05/,+1d' | tac > pdb_seqres_fixed.txt

Dominik

[ZiyaoLi]

Copy that. It looks that the script mishandled the DNA chains as fastas so that 0 is invalid. We are looking into the solution.

[ZiyaoLi]

Glad to see this solved, and thank you for the update. I believe they do not include dna & rna in the pdb_seqres.txt now, which is the reason of the problem.

[ZiyaoLi]

It seems that the parsing of file pdb_seqres.txt failed:

Parse failed (sequence file /home/data/pdb_seqres/pdb_seqres.txt):
Line 1360234: illegal character 0

Would you please upload the file you used? I understand that the file is big, so perhaps you can locate Line 1360234 and show the context nearby.

[ZiyaoLi]

Also @BaozCWJ may look into this.

[superantichrist]

7ooo_B mol:na length:11 DNA (5'-D(CPTP*(RWQ)PTPCPTPTPTPG)-3')
CT05ATCTTTG
7ooo_E mol:na length:11 DNA (5'-D(CPTP*(RWQ)PTPCPTPTPTPG)-3')
CT05ATCTTTG

second line is Line 1360234, and as error says, '0' character involved.

I used databases from alphafold2 reference site, and it works with their code.

[BaozCWJ]

7ooo_B mol:na length:11 DNA (5'-D(_CP_TP*(RWQ)P_TP_CP_TP_TP_TP_G)-3')
CT05ATCTTTG
7ooo_E mol:na length:11 DNA (5'-D(_CP_TP*(RWQ)P_TP_CP_TP_TP_TP_G)-3')
CT05ATCTTTG

second line is Line 1360234, and as error says, '0' character involved.

I used databases from alphafold2 reference site, and it works with their code.

It looks like the error of hmmsearch command
/usr/bin/hmmsearch --noali --cpu 8 --F1 0.1 --F2 0.1 --F3 0.1 --incE 100 -E 100 --domE 100 --incdomE 100 -A /tmp/tmpil8uqp24/output.sto /tmp/tmpil8uqp24/query.hmm /home/data/pdb_seqres/pdb_seqres.txt
could you please try the same fasta with alphafold's code and upload the coressponding hmmsearch command?

[superantichrist]

I checked alphafold code and It appears hmmsearch only used in multimer model, and when I run multimer, same error occured.

command seems same. (see below)
I0809 13:25:16.533803 140494127470400 run_docker.py:255] I0809 04:25:16.531341 139681435416384 hmmsearch.py:103] Launching sub-process ['/usr/bin/hmmsearch', '--noali', '--cpu', '8', '--F1', '0.1', '--F2', '0.1', '--F3', '0.1', '--incE', '100', '-E', '100', '--domE', '100', '--incdomE', '100', '-A', '/tmp/tmp69wsql5x/output.sto', '/tmp/tmp69wsql5x/query.hmm', '/mnt/pdb_seqres_database_path/pdb_seqres.txt']
....(omitted)
I0809 13:25:28.802923 140494127470400 run_docker.py:255] stderr:
I0809 13:25:28.803056 140494127470400 run_docker.py:255] Parse failed (sequence file /mnt/pdb_seqres_database_path/pdb_seqres.txt):
I0809 13:25:28.803190 140494127470400 run_docker.py:255] Line 1360234: illegal character 0

Sorry for inaccurate info that alphafold2 ref works.

So, Is my pdb_seqres.txt file a problem?

[superantichrist]

ftp://ftp.wwpdb.org/pub/pdb/derived_data/pdb_seqres.txt seems modified. I checked my old pdb_seqres.txt and find differences.
I re-run with old one and error not occured!
Thanks for your reply.

[dominik-handler]

Hi, I ran into the same problem. Is there a fix that works with the new PDB_seqres version? Or alternatively, do you know if the old file is still available somewhere?

Thank you,
Dominik

[ZiyaoLi]

@dominik-handler Thank you for the report, and the hot-fix for this issue. Since it's been reported twice, we are now looking into a solution to automatically fix this issue.

[dominik-handler]

Thank you!
By the way the database file-structure created by download_all does not fit the expected file naming of run_unifold.sh.
I had to manually fix the structure of the uniprot database.

[ZiyaoLi]

Thank you! By the way the database file-structure created by download_all does not fit the expected file naming of run_unifold.sh. I had to manually fix the structure of the uniprot database.

Maybe @BaozCWJ can look into this?

[BaozCWJ]

Thank you!
By the way the database file-structure created by download_all does not fit the expected file naming of run_unifold.sh.
I had to manually fix the structure of the uniprot database.

Thanks for your issue, could you please confirm the correct path of uniprot database as following?
--uniprot_database_path=$database_dir/uniprot/uniprot.fasta

[dominik-handler]

That should be correct. Thank you!

[BaozCWJ]

That should be correct. Thank you!

Thx! It's already fixed in the #42