Substitutions with an equivalent property recommended at a same position

[GuillaumeH-GR]

Hello,

First, thank you for efficient-evolution and congrats for your recent paper !

During an evaluation, I noticed that the same "type" of substitution was recommended several times at the same position. By this, I mean that different amino acids of equivalent property were recommended at the same position. One example for the VL of my antibody of interest : H101L (2), H101I (1), H101F (1) - with the number of models between brackets. Three large or very large hydrophobic amino acids recommended at the same position.

You used the number of language models as a criterion for choosing the substitutions to test experimentally but did you try any other ? Like, in this case, a "same" property selected several times at one position?

Thank you for your consideration,

Guillaume H