Quantitative structure-activity relationship (QSAR) models trained on a single data source tend to have limited coverage against data from other sources. This is a particular issue when predicting proprietary data using models trained on public data. Presented at the SOT, this poster looks at Effiris, a consortium-led federated QSAR model for secondary pharmacology, with particular attention on data preparation.
Consortium-led federated QSAR models for secondary pharmacology – preparing the data
Davies R; Williams RV; Fowkes A; Johnston LA;