other organisms?

[aminakur]

Can this tool be implemented for non-human genomes? If not, what modifications to the source code need to be made for the use of custom genomes?

[makirc]

In principle yes. Practically, that is a lot of work. You need to collect relevant genome-wide annotations from the other species and most importantly you need to collect a very unbiased (i.e. with regard to genomic location, size and annotations) set of benign SVs that you can match with randomly drawn events of the same size. To scale the scores, it would be useful to have a large population SV set for which you can report relative ranking of the scores. Once you have all of that, you can train a new model and do new predictions for other SVs.

People have done these things for CADD, but as I said, it is a lot of effort. See publications here:
https://pubmed.ncbi.nlm.nih.gov/32033531/ (CADD for Pigs)
https://www.biorxiv.org/content/10.1101/2024.11.01.621569v1.full.pdf (CADD for two birds)