AB-Amy is the first tool for predicting AL amyloidosis in therapeutic antibodies, which enables rapid screening of antibody drug candidates with the risk of amyloidosis in early development stages, thereby saving cost and time in drug development.
In this study, we collected 742 amyloidogenic and 712 non-amyloidogenic VL sequences to build a support vector machine -based model, named AB-Amy, for predicting the amyloidosis tendency of therapeutic antibodies. AB-Amy achieved the sensitivity, specificity, MCC, ACC and AUC of 93.80%, 91.98%, 92.95%, 0.8584 and 0.9651, respectively.
Citation:Yuwei Zhou and others, AB-Amy: machine learning aided amyloidogenic risk prediction of therapeutic antibody light chains, Antibody Therapeutics, 2023;, tbad007, https://doi.org/10.1093/abt/tbad007