option to define a3m file name based on protein header?

[flyark]

I am currently using 1.5.2 localcolabfold. After executing colabfold_search, I found that all the a3m file names are numerical.

0.a3m
1.a3m
2.a3m
...
x.a3m

Inside of the individual a3m file, I can see this via "head" in the terminal in order to see the protein name only.

ak $ head *.a3m -n 2
==> 0.a3m <==
#124,403	1,1
>Protein_1	Protein_1

==> 2.a3m <==
#124,729	1,1
>Protein_3	Protein_3

My fasta file used for sequence search looks like this.

>Protein_1 & Protein_2
MTKTKGEKINKSAINEVVTRECTIHLAKRVHNIGFKKRAPRAIKEIRKFAEREMGTTDVRIDTRLNKHIWSKGIRSTPFRIRVRLARRRNDDEDSPNKLYTYVTYVPVSTFKNLQTENVESSDD:MDNSGNNRYELLFM
>Protein_3 & Protein_4
MVKVKCSELRIKDKKELTKQLDELKNELLSLRVAKVTGGAPSKLSKIRVVRKAIARVYIVMHQKQKENLRKVFKNKKYKPLDLRKKKTRAIRKALSPRDANRKTLKEIRKRSVFPQRKFAVKA:MQLAEKHIVALMFPAIDVSTFTFVSGVKFYIMDNSGNNRYELLFMDDDDSSGLAQPQIAAVVAAPKKPEPAKAPKAPKSKSEKENKPVVAARKANAPVAKNASPVKGG

When protein header in the fasta file contains &, it seems only first protein/gene name is used for saving protein header in a3m file.

If I use the fasta file for MSA via colabfold_batch (remote server), I can get file names like the below.
Protein_1___Protein_2.a3m
Protein_3___Protein_4.a3m

Is there a way to have file name like this when colabfold_search is used for MSA?

And is there example fasta file for multimer analysis? If protein headers are unusual, I can refer to the example fasta file.

[huiwenke]

You can use MineProt toolkit (https://github.com/huiwenke/MineProt/wiki/Toolkit-manual#colabfoldtransformpy) to rename them.