Wrong in single sequence mode

[quailwwk]

Hi, I tried to run AlphaFold2.ipynb in single sequence mode, but the result is completely same to MMseq2 (UniRef only) mode.

Sequence:

GSHMSSVTIGKCYIQNRENGGRAFYNLGRKDLGIFTGKMYDDQIWSFQKSDTPGYYTIGRESKFLQYNGEQVIMSDIEQDTTLWSLEEVPEDKGFYRLLNKVHKAYLDYNGGDLVANKHQTESEKWILFKAY

config:

# single sequence mode
num_models=1
num_recycles=1
use_amber=False
use_msa=False
msa_mode=single_sequence
use_templates=False

# MMseq2 (UniRef only) mode
num_models=1
num_recycles=1
use_amber=False
use_msa=True
msa_mode=MMseqs2 (UniRef only)
use_templates=False

The top rows of pdb predicted by single-sequnce mode: MODEL 1
ATOM 1 N GLY A 1 -16.476 23.049 -12.538 1.00 30.08 N
ATOM 2 CA GLY A 1 -17.257 22.145 -13.366 1.00 30.08 C
ATOM 3 C GLY A 1 -16.750 20.715 -13.331 1.00 30.08 C
ATOM 4 O GLY A 1 -15.586 20.456 -13.644 1.00 30.08 O

The top rows of pdb predicted by MMseqs mode: MODEL 1
ATOM 1 N GLY A 1 -16.476 23.049 -12.538 1.00 30.08 N
ATOM 2 CA GLY A 1 -17.257 22.145 -13.366 1.00 30.08 C
ATOM 3 C GLY A 1 -16.750 20.715 -13.331 1.00 30.08 C
ATOM 4 O GLY A 1 -15.586 20.456 -13.644 1.00 30.08 O

[sokrypton]

How many sequences were found by MMseqs2? If none, than the answer should be identical.

[martin-steinegger]

Thank you for reporting this. I can confirm that "single_sequence" is currently not work. Will fix it soon.

[quailwwk]

~70 xxx_uniref.a3m

How many sequences were found by MMseqs2? If none, than the answer should be identical.

[martin-steinegger]

@quailwwk this should be resolved now. If you still have the same active session then you need to do a factory reset.