search

YoshitakaMo / localcolabfold

ColabFold on your local PC
MIT License
562 stars 129 forks source link

Could not get MSA/templates (HHSearch failed) #107

Closed agatawitkowska closed 1 year ago

agatawitkowska commented 2 years ago

On Windows 10 using WSL I'm trying to run:

~$ colabfold_batch --amber --templates --num-recycle 3 --use-gpu-relax --model-type AlphaFold2-multimer-v2 PDBID_6IWD.csv outputdir/ --cpu

using PDBID_6IWD.csv:

id,sequence
PDBID_6IWD,MVDATRVPMDERFRTLKKKLEEGMVFTEYEQIPKKKANGIFSTAALPENAERSRIREVVPYEENRVELIPTKENNTGYINASHIKVVVGGAEWHYIATQGPLPHTCHDFWQMVWEQGVNVIAMVTAEEEGGRTKSHRYWPKLGSKHSSATYGKFKVTTKFRTDSVCYATTGLKVKHLLSGQERTVWHLQYTDWPDHGCPEDVQGFLSYLEEIQSVRRHTNSMLEGTKNRHPPIVVHCSAGVGRTGVLILSELMIYCLEHNEKVEVPMMLRLLREQRMFMIQTIAQYKFVYQVLIQFLQNSRLI:MVDATRVPMDERFRTLKKKLEEGMVFTEYEQIPKKKANGIFSTAALPENAERSRIREVVPYEENRVELIPTKENNTGYINASHIKVVVGGAEWHYIATQGPLPHTCHDFWQMVWEQGVNVIAMVTAEEEGGRTKSHRYWPKLGSKHSSATYGKFKVTTKFRTDSVCYATTGLKVKHLLSGQERTVWHLQYTDWPDHGCPEDVQGFLSYLEEIQSVRRHTNSMLEGTKNRHPPIVVHCSAGVGRTGVLILSELMIYCLEHNEKVEVPMMLRLLREQRMFMIQTIAQYKFVYQVLIQFLQNSRLI:GHMAEPQRHTMLCMCCKCEARIELVVESSADDLRAFQQLFLNTLSFVCPWCASQQ:GHMAEPQRHTMLCMCCKCEARIELVVESSADDLRAFQQLFLNTLSFVCPWCASQQ

I'm getting:

2022-08-19 14:14:29,095 Running colabfold 1.3.0
2022-08-19 14:14:29,781 --max-msa can not be used in combination with AlphaFold2-multimer (--max-msa ignored)
WARNING: You are welcome to use the default MSA server, however keep in mind that it's a limited shared resource only capable of processing a few thousand MSAs per day. Please submit jobs only from a single IP address. We reserve the right to limit access to the server case-by-case when usage exceeds fair use.

If you require more MSAs:

* You can precompute all MSAs with `colabfold_search` or

* You can host your own API and pass it to `--host-url`
/home/username/.local/lib/python3.8/site-packages/haiku/_src/data_structures.py:37: FutureWarning: jax.tree_structure is deprecated, and will be removed in a future release. Use jax.tree_util.tree_structure instead.
  PyTreeDef = type(jax.tree_structure(None))
2022-08-19 14:14:32,042 Found 8 citations for tools or databases
2022-08-19 14:14:41,237 Query 1/1: PDBID_6IWD (length 716)
2022-08-19 14:14:42,611 Sleeping for 8s. Reason: PENDING
2022-08-19 14:14:51,872 Sleeping for 9s. Reason: RUNNING
2022-08-19 14:15:02,139 Sleeping for 10s. Reason: RUNNING
2022-08-19 14:15:13,409 Sleeping for 10s. Reason: RUNNING
2022-08-19 14:15:24,672 Sleeping for 9s. Reason: RUNNING
2022-08-19 14:15:34,928 Sleeping for 10s. Reason: RUNNING
COMPLETE: 100%|██████████████████████████████████████████████████████████████| 300/300 [elapsed: 01:16 remaining: 00:00]
2022-08-19 14:16:22,815 Could not get MSA/templates for PDBID_6IWD: HHSearch failed:
stdout:

stderr:
- 14:16:20.510 INFO: /tmp/tmp3df2iks3/query.a3m is in A2M, A3M or FASTA format

- 14:16:20.684 INFO: Searching 20 database HHMs without prefiltering

- 14:16:20.684 INFO: Iteration 1

- 14:16:20.813 WARNING: database contains sequences that exceeds maximum allowed size (maxres = 20001). Maxres can be increased with parameter -maxres.

- 14:16:20.831 INFO: Scoring 20 HMMs using HMM-HMM Viterbi alignment

- 14:16:20.858 INFO: Alternative alignment: 0

- 14:16:21.276 INFO: 20 alignments done

- 14:16:21.276 INFO: Alternative alignment: 1

- 14:16:21.691 INFO: 20 alignments done

- 14:16:21.691 INFO: Alternative alignment: 2

- 14:16:21.921 INFO: 8 alignments done

- 14:16:21.921 INFO: Alternative alignment: 3

- 14:16:21.935 INFO: Realigning 28 HMM-HMM alignments using Maximum Accuracy algorithm

- 14:16:22.787 INFO: 28 sequences belonging to 28 database HMMs found with an E-value < 0.001

free(): invalid pointer

Traceback (most recent call last):
  File "/home/username/.local/lib/python3.8/site-packages/colabfold/batch.py", line 1332, in run
    ) = get_msa_and_templates(
  File "/home/username/.local/lib/python3.8/site-packages/colabfold/batch.py", line 787, in get_msa_and_templates
    template_feature = mk_template(
  File "/home/username/.local/lib/python3.8/site-packages/colabfold/batch.py", line 165, in mk_template
    hhsearch_result = hhsearch_pdb70_runner.query(a3m_lines)
  File "/home/username/.local/lib/python3.8/site-packages/alphafold/data/tools/hhsearch.py", line 94, in query
    raise RuntimeError(
RuntimeError: HHSearch failed:
stdout:

stderr:
- 14:16:20.510 INFO: /tmp/tmp3df2iks3/query.a3m is in A2M, A3M or FASTA format

- 14:16:20.684 INFO: Searching 20 database HHMs without prefiltering

- 14:16:20.684 INFO: Iteration 1

- 14:16:20.813 WARNING: database contains sequences that exceeds maximum allowed size (maxres = 20001). Maxres can be increased with parameter -maxres.

- 14:16:20.831 INFO: Scoring 20 HMMs using HMM-HMM Viterbi alignment

- 14:16:20.858 INFO: Alternative alignment: 0

- 14:16:21.276 INFO: 20 alignments done

- 14:16:21.276 INFO: Alternative alignment: 1

- 14:16:21.691 INFO: 20 alignments done

- 14:16:21.691 INFO: Alternative alignment: 2

- 14:16:21.921 INFO: 8 alignments done

- 14:16:21.921 INFO: Alternative alignment: 3

- 14:16:21.935 INFO: Realigning 28 HMM-HMM alignments using Maximum Accuracy algorithm

- 14:16:22.787 INFO: 28 sequences belonging to 28 database HMMs found with an E-value < 0.001

free(): invalid pointer

2022-08-19 14:16:22,816 Done

I would appreciate any helpful advice.

biotech70 commented 2 years ago

can you share with me your CSV file?

agatawitkowska commented 2 years ago

I pasted content of PDBID_6IWD.csv above (that's one I tried). Now also as a file: PDBID_6IWD.csv.

biotech70 commented 2 years ago

its work for me

agatawitkowska commented 2 years ago

Sure. I do not expect problems with my PDBID_6IWD.csv or run command. Thanks anyway.

I guess there might be some "install"/dependencies problem which I'm not sure how to start debugging right now.

lukaszjablonski commented 2 years ago

@biotech70, did you run it on Windows 10 using WSL Ubuntu 20.04.4 LTS (GNU/Linux 5.10.102.1-microsoft-standard-WSL2 x86_64)?

biotech70 commented 2 years ago

no

IlyesAbdelhamid commented 1 year ago

Dear all,

I've been using ColabFold for thousands of pairs of sequences and for the first time I have the same problem as the author of this issue.

This is my prompt: colabfold_batch --num-recycle 3 --templates --pair-mode unpaired+paired --model-type AlphaFold2-multimer-v2 --rank intscore $SEQUENCE_FILE$PAIR_TO_TEST $COLABFOLD_OUTPUT$PAIR_TO_TEST

You will find below the fasta sequence of the pair of proteins:

P43132 MKLGIIPYQEGTDIVYKNALQGQQEGKRPNLPQMEATHQIKSSVQGTSYEFVRTEDIPLNRRHFVYRPCS ANPFFTILGYGCTEYPFDHSGMSVMDRSEGLSISRDGNDLVSVPDQYGWRTARSDVCIKEGMTYWEVEVI RGGNKKFADGVNNKENADDSVDEVQSGIYEKMHKQVNDTPHLRFGVCRREASLEAPVGFDVYGYGIRDIS LESIHEGKLNCVLENGSPLKEGDKIGFLLSLPSIHTQIKQAKEFTKRRIFALNSHMDTMNEPWREDAENG PSRKKLKQETTNKEFQRALLEDIEYNDVVRDQIAIRYKNQLFFEATDYVKTTKPEYYSSDKRERQDYYQL EDSYLAIFQNGKYLGKAFENLKPLLPPFSELQYNEKFYLGYWQHGEARDESNDKNTTSAKKKKQQQKKKK GLILRNKYVNNNKLGYYPTISCFNGGTARIISEEDKLEYLDQIRSAYCVDGNSKVNTLDTLYKEQIAEDI VWDIIDELEQIALQQ

P39706 MNILLQDPFAVLKEHPEKLTHTIENPLRTECLQFSPCGDYLALGCANGALVIYDMDTFRPICVPGNMLGA HVRPITSIAWSPDGRLLLTSSRDWSIKLWDLSKPSKPLKEIRFDSPIWGCQWLDAKRRLCVATIFEESDA YVIDFSNDPVASLLSKSDEKQLSSTPDHGYVLVCTVHTKHPNIIIVGTSKGWLDFYKFHSLYQTECIHSL KITSSNIKHLIVSQNGERLAINCSDRTIRQYEISIDDENSAVELTLEHKYQDVINKLQWNCILFSNNTAE YLVASTHGSSAHELYIWETTSGTLVRVLEGAEEELIDINWDFYSMSIVSNGFESGNVYVWSVVIPPKWSA LAPDFEEVEENVDYLEKEDEFDEVDEAEQQQGLEQEEEIAIDLRTREQYDVRGNNLLVERFTIPTDYTRI IKMQSS

And you will find below the output file of my job in the SLURM cluster.

The following have been reloaded with a version change: 1) modenv/scs5 => modenv/hiera

Module GCC/10.2.0CUDA/11.3.1OpenMPI/4.0.5 and 12 dependencies loaded. sbatch run_colabfold_AF2Mv2_unpaired_paired_net_perturbed.sh 1 Submitted batch job 29707500 WARNING: You are welcome to use the default MSA server, however keep in mind that it's a limited shared resource only capable of processing a few thousand MSAs per day. Please submit jobs only from a single IP address. We reserve the right to limit access to the server case-by-case when usage exceeds fair use.

If you require more MSAs:

stderr:

I purposefully truncated the rest of the output because it goes on the same over thousands lines. Thank you in advance for your help.

Sincerely, Ilyes

biotech70 commented 1 year ago

please share the provided seq file that was used for structure prediction.

IlyesAbdelhamid commented 1 year ago

You will find the seq file in the attached document.

Sincerely, Ilyes

Le jeu. 6 oct. 2022 à 16:12, mohammad mahmoudi gomari < @.***> a écrit :

please share the provided seq file that was used for structure prediction.

— Reply to this email directly, view it on GitHub https://github.com/YoshitakaMo/localcolabfold/issues/107#issuecomment-1270126498, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEKMR56BQULUTL6HMNLF42DWB3M3TANCNFSM57ASAVDA . You are receiving this because you commented.Message ID: @.***>

biotech70 commented 1 year ago

for multimer prediction, you must separate desired sequences by using the : symbol in a CSV file.

biotech70 commented 1 year ago

like below: MKLGIIPYQEGTDIVYKNALQGQQEGKRPNLPQMEATHQIKSSVQGTSYEFVRTEDIPLNRRHFVYRPCS ANPFFTILGYGCTEYPFDHSGMSVMDRSEGLSISRDGNDLVSVPDQYGWRTARSDVCIKEGMTYWEVEVI RGGNKKFADGVNNKENADDSVDEVQSGIYEKMHKQVNDTPHLRFGVCRREASLEAPVGFDVYGYGIRDIS LESIHEGKLNCVLENGSPLKEGDKIGFLLSLPSIHTQIKQAKEFTKRRIFALNSHMDTMNEPWREDAENG PSRKKLKQETTNKEFQRALLEDIEYNDVVRDQIAIRYKNQLFFEATDYVKTTKPEYYSSDKRERQDYYQL EDSYLAIFQNGKYLGKAFENLKPLLPPFSELQYNEKFYLGYWQHGEARDESNDKNTTSAKKKKQQQKKKK GLILRNKYVNNNKLGYYPTISCFNGGTARIISEEDKLEYLDQIRSAYCVDGNSKVNTLDTLYKEQIAEDI VWDIIDELEQIALQQ:MNILLQDPFAVLKEHPEKLTHTIENPLRTECLQFSPCGDYLALGCANGALVIYDMDTFRPICVPGNMLGA HVRPITSIAWSPDGRLLLTSSRDWSIKLWDLSKPSKPLKEIRFDSPIWGCQWLDAKRRLCVATIFEESDA YVIDFSNDPVASLLSKSDEKQLSSTPDHGYVLVCTVHTKHPNIIIVGTSKGWLDFYKFHSLYQTECIHSL KITSSNIKHLIVSQNGERLAINCSDRTIRQYEISIDDENSAVELTLEHKYQDVINKLQWNCILFSNNTAE YLVASTHGSSAHELYIWETTSGTLVRVLEGAEEELIDINWDFYSMSIVSNGFESGNVYVWSVVIPPKWSA LAPDFEEVEENVDYLEKEDEFDEVDEAEQQQGLEQEEEIAIDLRTREQYDVRGNNLLVERFTIPTDYTRI IKMQSS

IlyesAbdelhamid commented 1 year ago

I forgot to mention that I modified the code internally to get the same required format as you showed above.

Sincerely, Ilyes

Le jeu. 6 oct. 2022 à 16:32, ilyes abdelhamid @.***> a écrit :

I've been computing multimer prediction with the format I sent you in the fasta file.

Sincerely, Ilyes

Le jeu. 6 oct. 2022 à 16:23, mohammad mahmoudi gomari < @.***> a écrit :

like below: MKLGIIPYQEGTDIVYKNALQGQQEGKRPNLPQMEATHQIKSSVQGTSYEFVRTEDIPLNRRHFVYRPCS ANPFFTILGYGCTEYPFDHSGMSVMDRSEGLSISRDGNDLVSVPDQYGWRTARSDVCIKEGMTYWEVEVI RGGNKKFADGVNNKENADDSVDEVQSGIYEKMHKQVNDTPHLRFGVCRREASLEAPVGFDVYGYGIRDIS LESIHEGKLNCVLENGSPLKEGDKIGFLLSLPSIHTQIKQAKEFTKRRIFALNSHMDTMNEPWREDAENG PSRKKLKQETTNKEFQRALLEDIEYNDVVRDQIAIRYKNQLFFEATDYVKTTKPEYYSSDKRERQDYYQL EDSYLAIFQNGKYLGKAFENLKPLLPPFSELQYNEKFYLGYWQHGEARDESNDKNTTSAKKKKQQQKKKK GLILRNKYVNNNKLGYYPTISCFNGGTARIISEEDKLEYLDQIRSAYCVDGNSKVNTLDTLYKEQIAEDI

VWDIIDELEQIALQQ:MNILLQDPFAVLKEHPEKLTHTIENPLRTECLQFSPCGDYLALGCANGALVIYDMDTFRPICVPGNMLGA HVRPITSIAWSPDGRLLLTSSRDWSIKLWDLSKPSKPLKEIRFDSPIWGCQWLDAKRRLCVATIFEESDA YVIDFSNDPVASLLSKSDEKQLSSTPDHGYVLVCTVHTKHPNIIIVGTSKGWLDFYKFHSLYQTECIHSL KITSSNIKHLIVSQNGERLAINCSDRTIRQYEISIDDENSAVELTLEHKYQDVINKLQWNCILFSNNTAE YLVASTHGSSAHELYIWETTSGTLVRVLEGAEEELIDINWDFYSMSIVSNGFESGNVYVWSVVIPPKWSA LAPDFEEVEENVDYLEKEDEFDEVDEAEQQQGLEQEEEIAIDLRTREQYDVRGNNLLVERFTIPTDYTRI IKMQSS

— Reply to this email directly, view it on GitHub https://github.com/YoshitakaMo/localcolabfold/issues/107#issuecomment-1270148430, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEKMR52IUAXXRVZO5LBYWPLWB3OHXANCNFSM57ASAVDA . You are receiving this because you commented.Message ID: @.***>

biotech70 commented 1 year ago

I don't understand, do you get an error when you use a CSV file where the sequences are separated by : symbol?

biotech70 commented 1 year ago

please use from attached file and below command and let me know the output: colabfold_batch --num-recycle 3 --templates --model-type AlphaFold2-multimer-v2 --rank intscore 6WOV.csv outputdir/ 6WOV.csv

biotech70 commented 1 year ago

please provide information on your OS and GPU

IlyesAbdelhamid commented 1 year ago

You will find in the attached document the output of ColabFold. It is the same problem as the one I initially reported: Could not get MSA/templates for PDBID_6WOV: HHSearch failed

I use a multi-GPU sub-cluster "Alpha Centauri".. It has 34 nodes, each with:

Le jeu. 6 oct. 2022 à 17:10, mohammad mahmoudi gomari < @.***> a écrit :

please provide information on your OS and GPU

— Reply to this email directly, view it on GitHub https://github.com/YoshitakaMo/localcolabfold/issues/107#issuecomment-1270235420, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEKMR54HYKINIW3HY3MYOCDWB3TUVANCNFSM57ASAVDA . You are receiving this because you commented.Message ID: @.***>

IlyesAbdelhamid commented 1 year ago

I've been computing multimer prediction with the format I sent you in the fasta file.

Sincerely, Ilyes

Le jeu. 6 oct. 2022 à 16:23, mohammad mahmoudi gomari < @.***> a écrit :

like below: MKLGIIPYQEGTDIVYKNALQGQQEGKRPNLPQMEATHQIKSSVQGTSYEFVRTEDIPLNRRHFVYRPCS ANPFFTILGYGCTEYPFDHSGMSVMDRSEGLSISRDGNDLVSVPDQYGWRTARSDVCIKEGMTYWEVEVI RGGNKKFADGVNNKENADDSVDEVQSGIYEKMHKQVNDTPHLRFGVCRREASLEAPVGFDVYGYGIRDIS LESIHEGKLNCVLENGSPLKEGDKIGFLLSLPSIHTQIKQAKEFTKRRIFALNSHMDTMNEPWREDAENG PSRKKLKQETTNKEFQRALLEDIEYNDVVRDQIAIRYKNQLFFEATDYVKTTKPEYYSSDKRERQDYYQL EDSYLAIFQNGKYLGKAFENLKPLLPPFSELQYNEKFYLGYWQHGEARDESNDKNTTSAKKKKQQQKKKK GLILRNKYVNNNKLGYYPTISCFNGGTARIISEEDKLEYLDQIRSAYCVDGNSKVNTLDTLYKEQIAEDI

VWDIIDELEQIALQQ:MNILLQDPFAVLKEHPEKLTHTIENPLRTECLQFSPCGDYLALGCANGALVIYDMDTFRPICVPGNMLGA HVRPITSIAWSPDGRLLLTSSRDWSIKLWDLSKPSKPLKEIRFDSPIWGCQWLDAKRRLCVATIFEESDA YVIDFSNDPVASLLSKSDEKQLSSTPDHGYVLVCTVHTKHPNIIIVGTSKGWLDFYKFHSLYQTECIHSL KITSSNIKHLIVSQNGERLAINCSDRTIRQYEISIDDENSAVELTLEHKYQDVINKLQWNCILFSNNTAE YLVASTHGSSAHELYIWETTSGTLVRVLEGAEEELIDINWDFYSMSIVSNGFESGNVYVWSVVIPPKWSA LAPDFEEVEENVDYLEKEDEFDEVDEAEQQQGLEQEEEIAIDLRTREQYDVRGNNLLVERFTIPTDYTRI IKMQSS

— Reply to this email directly, view it on GitHub https://github.com/YoshitakaMo/localcolabfold/issues/107#issuecomment-1270148430, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEKMR52IUAXXRVZO5LBYWPLWB3OHXANCNFSM57ASAVDA . You are receiving this because you commented.Message ID: @.***>

biotech70 commented 1 year ago

oh so good

IlyesAbdelhamid commented 1 year ago

Do you have any idea what could be the issue ?

Sincerely, Ilyes

Le mar. 11 oct. 2022 à 10:13, mohammad mahmoudi gomari < @.***> a écrit :

oh so good

— Reply to this email directly, view it on GitHub https://github.com/YoshitakaMo/localcolabfold/issues/107#issuecomment-1274279876, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEKMR5YYF7AOGZKDVOEXF4TWCUOTFANCNFSM57ASAVDA . You are receiving this because you commented.Message ID: @.***>

biotech70 commented 1 year ago

It may be related to the headers of your sequences.

IlyesAbdelhamid commented 1 year ago

The headers ? From 6WOV.csv https://github.com/YoshitakaMo/localcolabfold/files/9725935/6WOV.csv ? Could you clarify please ?

Sincerely, Ilyes

Le mar. 11 oct. 2022 à 13:00, mohammad mahmoudi gomari < @.***> a écrit :

It may be related to the headers of your sequences.

— Reply to this email directly, view it on GitHub https://github.com/YoshitakaMo/localcolabfold/issues/107#issuecomment-1274505893, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEKMR575HTR5EJY7QBFVAZ3WCVCGBANCNFSM57ASAVDA . You are receiving this because you commented.Message ID: @.***>

biotech70 commented 1 year ago

Did you use the CSV file which I shared with you to determine the structure?

IlyesAbdelhamid commented 1 year ago

Yes. Please see the output in the attached document.

Sincerely, Ilyes

Le mar. 11 oct. 2022 à 13:10, mohammad mahmoudi gomari < @.***> a écrit :

Did you use the CSV file which I shared with you to determine the structure?

— Reply to this email directly, view it on GitHub https://github.com/YoshitakaMo/localcolabfold/issues/107#issuecomment-1274517056, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEKMR57NFMBI2XOIE6HNG3LWCVDMFANCNFSM57ASAVDA . You are receiving this because you commented.Message ID: @.***>

biotech70 commented 1 year ago

Ok, the most crucial option in prediction of structures by a CSV file is to use : symbol for the separation of sequences in multimeric form.

IlyesAbdelhamid commented 1 year ago

Yes the ':' is in the CSV file you sent to me. Please if you have any idea what could be the issue let me know. Thank you!

Sincerely, Ilyes

Le mar. 11 oct. 2022 à 13:27, mohammad mahmoudi gomari < @.***> a écrit :

Ok, the most crucial option in prediction of structures by a CSV file is to use : symbol for the separation of sequences in multimeric form.

— Reply to this email directly, view it on GitHub https://github.com/YoshitakaMo/localcolabfold/issues/107#issuecomment-1274534079, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEKMR54JQQ5YKOT7NYQMTJTWCVFKPANCNFSM57ASAVDA . You are receiving this because you commented.Message ID: @.***>

biotech70 commented 1 year ago

I listed below some things that are important in prediction using local-colabfold: 1- JAX version 2- GPU config 3- The size of the target sequence 4- The version of local-colabfold (because some people get different output from the old version of colabfold compared to the newer version of colabfold)

biotech70 commented 1 year ago

Finally, we should thank Dr. Yoshitaka Moriwaki, who made it possible to run colabfold on local machines.

IlyesAbdelhamid commented 1 year ago

Thank you, see reply below.

Sincerely, Ilyes

---------- Forwarded message --------- De : mohammad mahmoudi gomari @.> Date: mar. 11 oct. 2022 à 13:52 Subject: Re: [YoshitakaMo/localcolabfold] Could not get MSA/templates (HHSearch failed) (Issue #107) To: YoshitakaMo/localcolabfold @.> Cc: IlyesAbdelhamid @.>, Comment < @.>

I listed below some things that are important in prediction using local-colabfold: 1- JAX version 2- GPU config

REPLY: I hardly believe that the issue is due to JAX and GPU config because I've been predicting thousands of structures no problem. It's the only time I have such an issue.

3- The size of the target sequence

REPLY: Protein A is 505aa long and Protein B is 426aa long. I have dealt with pairs of proteins with much longer sequences with no problem.

4- The version of local-colabfold (because some people get different output from the old version of colabfold compared to the newer version of colabfold)

REPLY: I would say it is the most plausible option. I use colabfold 1.3.0. Have you tried to run my prediction with my parameters on your own local-colabfold version as a sanity check ? Command line: colabfold_batch --num-recycle 3 --templates --model-type AlphaFold2-multimer-v2 --rank intscore 6WOV.csv outputdir/

Le mar. 11 oct. 2022 à 13:55, mohammad mahmoudi gomari < @.***> a écrit :

Finally, we should thank Dr. Yoshitaka Moriwaki, who made it possible to run colabfold on local machines.

— Reply to this email directly, view it on GitHub https://github.com/YoshitakaMo/localcolabfold/issues/107#issuecomment-1274565251, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEKMR52MAYLZEVUCUQWHBLTWCVIUXANCNFSM57ASAVDA . You are receiving this because you commented.Message ID: @.***>

lukaszjablonski commented 1 year ago

Actually, the original problem issued by @agatawitkowska was never solved and got no attention here.