Malay Bhattacharyya

The paper predicts drug ADMET properties using molecular descriptors with accuracy competitive with deep learning models, outperforming state-of-the-art on 3 datasets.

The paper uses k-mer distributions to predict the geographical origin of SARS-CoV-2 sequences with over 90% accuracy by under-sampling imbalanced data.

The paper analyzes early SARS-CoV-2 genomic sequences to trace their temporal changes and finds that n-mer distributions can predict collection timelines with 78% accuracy.