DaVinci Institute

Massively-multitask regression models (MMRMs) have revolutionized activity prediction for drug discovery. MMRMs trained on millions of compounds and many thousands of assays can predict bioactivity with accuracy comparable to 4-concentration IC50 experiments. This report compares six MMRMs: pQSAR, Alchemite, MT-DNN, MetaNN, Macau and IMC. Models were trained by experts in each method, on identical sets of 159 kinase and 4276 diverse ChEMBL assays, employing realistically novel training/test set splits. Results were compared both qualitatively and with statistical rigor. Our use-case is imputing full bioactivity profiles for the very sparse compound collections on which the models were trained. MMRMs performed much better than the single-task random forest regression (ST-RFR) model. Five MMRMs train all models simultaneously, so must leave out test-set measurements from all assays to avoid leakage (here 25% of data), whereas one method trains models one-at-a-time, so only holds out test data for that assay (< 1% of data). Thus, all algorithms were compared both using 75/25 splits, and when possible, 99 + / < 1 splits. Many MMRM evaluations achieved similar accuracy when tested on the same split. However, when evaluated on 75/25 splits, all MMRMs performed much worse than when evaluated on 99 + / < 1% splits. Thus, while many MMRMs produce comparable final production models (trained on all the data), models that require 75/25 splits greatly underestimate the accuracy of the final models. While outstanding for imputations, MMRMs proved little better than ST-RFR for compounds very unlike the training collection. Thus, MMRMs are best for hit-finding, off-target, promiscuity, MoA, polypharmacology or drug-repurposing within the training collection. Since accuracy is not a deciding factor, other pros and cons of each method are also described.

To create belonging in medical education, Wyatt et al. argue that institutions must consider diversity, equity and inclusion (DEI) from a systems‐based perspective that integrates each of these components.

This paper describes an effort to develop a new comprehensive and consistent database of airfoil performance metrics that can assist in the aircraft conceptual and preliminary design process. Under this effort, almost 2,000 different airfoils have been analyzed under identical conditions and the aerodynamic performance of the airfoils compiled into a large database. The airfoil analysis was performed using the XFOIL analysis code, which combines a vortex panel potential flow model and an integral boundary layer algorithm. This paper discusses the approach used for developing this database, some of the challenges encountered during the process, and results of validation done to gauge the accuracy of the computed performance data.