To safely select drugs for a specific patient or discover safe and effective therapeutics, accurate and robust prediction of patient-specific responses to a new chemical compound are key. However, patient data are often too scarce to train a generalized machine-learning model.

In a paper published October 17 in the journal Nature Machine Intelligence, the CUNY researchers describe their new AI model -- called context-aware deconfounding autoencoder (CODE-AE) -- and how it can screen novel drug compounds to accurately predict efficacy in humans. In tests, it was also able to theoretically identify personalized drugs for over 9,000 patients that could better treat their conditions.

According to You Wu, a CUNY Graduate Center PhD student and co-author of the paper, CODE-AE can provide a workaround to the problem of having sufficient patient data to train a generalized machine-learning model.

"CODE-AE can extract intrinsic biological signals masked by noise and confounding factors and effectively alleviated the data-discrepancy problem," Wu said.

As a result, CODE-AE significantly improves accuracy and robustness with state-of-the-art methods in predicting patient-specific drug responses purely from cell-line compound screens, according to researchers.

Next, the researchers aim to develop a way for CODE-AE to reliably predict the effect of a new drug's concentration and metabolization in human bodies. They also noted that the AI model could potentially be tweaked to accurately predict human side effects to drugs.

The study was supported by the National Institute of General Medical Sciences and the National Institute on Aging.