Getting BLAST-like e-value for TM-align?

[meltzj]

Hi there,

Thank you so much for your constant work with foldseek and developing such fantastic tools!

I just wanted to ask why it is that the way the "e-value" is calculated is different for TMalign vs 3Di/AA?

I understand that for 3Di/AA the e-value is calculated similarly to e-values for other alignment tools such as blast, however for TM align it is the average of the TM score normalized by query and by target.

Is there a way to perform a global alignment (TM align) using foldseek and still getting a "BLAST-like" e-value in the output?

Thank you very much

Julia

[meltzj]

Actually I realise that you have developed a TM-score "evalue" that makes more sense in it's purpose in evaluating the randomness of the TM alignment than a "blast-like e-value" however what do you recommend as the benchmark TM evalue for evaluating randomness of TM-alignment?