Are Casanovo per-AA scores reversed?

[justin-a-sanders]

Describe the Issue

It looks like the list of per-amino-acid scores reported by Casanovo might be reversed. For most incorrect predictions, it looks like the lowest scoring amino acids would line up better with the sequencing errors if the list of scores was reversed. For example:

Casanovo peptide: TSSGGTAALGCLVK	
Tide peptide:     STSGGTAALGCLVK	
Per-AA scores: 0.98486,0.98473,0.98619,0.98608,0.98527,0.98520,0.98420,0.98520,0.98551,0.98500,0.98550,0.98627,0.91044,0.98314

Peptides are sequenced from right-to-left and then reversed to get the final prediction. Are the corresponding per-aa scores also reversed accordingly?

[bittremieux]

Can you please confirm the Casanovo version you're using?

[justin-a-sanders]

I was seeing this in results from Casanovo version 4.2.1