Calculation of amino acid probabilities

[diamondspark]

How were the probabilities for amino acids in the class BaseSequence() calculated? Particularly, how was prob_ACPhel computed?
Thank You!

[alexarnimueller]

the AA probabilities are pre-computed based on a literature analysis of anticancer peptides. They are dependent on the sequence position and are defined here:
https://github.com/alexarnimueller/modlAMP/blob/60039e8a02f55d0797a6f0e25eeb046ff59eaadb/modlamp/core.py#L125C15-L125C15

The publication from which they originated is the following one, with a visualisation in Fig.3:
DOI https://doi.org/10.1039/C6MD00376A

Hope this helps

[diamondspark]

Thank you. I understand this. I meant how were these probs. precomputed. Do you have the code for it? I want to recompute with updated literature on anticancer peptides.

[alexarnimueller]

Great to hear you want to update the AA probabilities.
This was done by @ggabernet outside of modlamp.
Briefly: ACPs were aligned and then for every position up to the first 18AAs, AAs were counted and returned as a fraction. That resulted in the numbers provided in https://github.com/alexarnimueller/modlAMP/blob/60039e8a02f55d0797a6f0e25eeb046ff59eaadb/modlamp/core.py#L125C15-L125C15
@ggabernet correct me if I'm wrong