TECHNOLOGY
to investigate how different genes are related to drug response , have been mostly done on a relatively small number of populations . So I think there are a lot of individual segments of the global population that have not been well represented in studies so far .”
Where AI stands to benefit pharmacogenomics , is in its ability to deal with really complex issues .
“ We ' ve seen examples in the last several years where AI based algorithms have been able to give us better predictions of appropriate drug usage and dose ”, says Streetman . “ A lot of the early algorithms that were used to identify doses were based on one or two genes . Data was mostly from Caucasian populations and those algorithms did reasonably well for Caucasians , but they were much less robust , much less predictive for African Americans and for other groups .
“ Since then , we ' ve been able to identify a lot of other genetic predictors and genetic factors that contribute to differences in response that are much more important for
78 February 2022