ClairS and ClairS-TO for other diseases

[ccbruels]

Thank you for all of your work on these pipelines!

I've run ClairS and ClairS-TO successfully on several samples. However, we are analyzing samples from healthy control individuals and individuals with other disorders - not cancer. I suspect that the models being trained on specific cancer mutation types may have an impact on our analyses.

Do you have any suggestions for running ClairS/ClairS-TO on from other disease tissues? Or plans to train models for non-cancer analyses?

Thanks!

[zhengzhenxian]

Hi,

To avoid introducing specific cancer mutation types, please use our models trained on synthetic samples (e.g., ont_r10_dorado_sup_5khz_ss, hifi_revio_ss, etc.). These models were trained without incorporating any real cancer data, making them more generic.

If for germline variant calling, please try Clair3.