Assessing the performance of implicit ensembling of neural networks for proteochemometric modeling with RandomNets
- 1Leiden Academic Centre for Drug Research, Einsteinweg 55, Leiden, 2333CC
- 2Leiden Institute of Advanced Computer Science, Leiden University, Einsteinweg 55, Leiden, Netherlands, 2333CC
Abstract
RandomNets, a recently introduced neural-network architecture that emulates the feature and sample bagging of Random Forests through an implicit, vectorised ensemble, has shown promising performance for quantitative structure–activity relationship (QSAR) modelling, but so far only on random data splits. Here we evaluate RandomNets in a proteochemometric (PCM) setting on two benchmarks: an original dataset (BtH) and a newly prepared dataset derived from ChEMBL 35, comprising 571,197 interactions across 382,493 unique compounds and 2,023 unique targets. Models were built from Morgan fingerprints together with additional physicochemical and protein descriptors, and assessed under both random and temporal splits. All models were trained within QSPRpred, enabling reproducible and benchmark-agnostic comparison against the best-performing PCM models from BtH, a deep neural network (DNN_PCM) and a Random Forest (RF_PCM). RandomNets with two layers gave the best overall performance, notably improving on the Matthews correlation coefficient (MCC) of the DNN baseline for the temporal split by 5.55% on the original dataset and 8.73% on the new dataset, while taking 2× less time to train on the original dataset and the same time on the new dataset. These results affirm the potential of RandomNets and extend it to PCM modelling.
Related
- PCAM: proteochemometric assay-aware modeling for robust bioactivity estimation and uncertainty quantification
- Chemist-aligned retrosynthesis by ensembling diverse inductive bias models
- Assessing the performance of coupled-trajectory schemes on full-dimensional two-state linear vibronic coupling models
- IMPLICIT LEARNING: REACTION OUTCOME PREDICTION FROM POSITIVE-ONLY DATA
- Polarization-insensitive broadband achromatic metalens with implicit pupil apodization and focal sharpening
BibTeX
@misc{jespers2026assessing,
title = {{Assessing the performance of implicit ensembling of neural networks for proteochemometric modeling with RandomNets}},
author = {Chiel Jespers and Olivier J. M. Béquignon and Mike Preuss and Gerard J. P. van Westen},
howpublished = {ChemRxiv},
year = {2026},
month = sep,
doi = {10.26434/chemrxiv.15008684/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008684/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Assessing the performance of implicit ensembling of neural networks for proteochemometric modeling with RandomNets
AU - Chiel Jespers
AU - Olivier J. M. Béquignon
AU - Mike Preuss
AU - Gerard J. P. van Westen
PY - 2026
DA - 2026/09/11
PB - ChemRxiv
DO - 10.26434/chemrxiv.15008684/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15008684/v1
AB - RandomNets, a recently introduced neural-network architecture that emulates the feature and sample bagging of Random Forests through an implicit, vectorised ensemble, has shown promising performance for quantitative structure–activity relationship (QSAR) modelling, but so far only on random data splits. Here we evaluate RandomNets in a proteochemometric (PCM) setting on two benchmarks: an original dataset (BtH) and a newly prepared dataset derived from ChEMBL 35, comprising 571,197 interactions across 382,493 unique compounds and 2,023 unique targets. Models were built from Morgan fingerprints together with additional physicochemical and protein descriptors, and assessed under both random and temporal splits. All models were trained within QSPRpred, enabling reproducible and benchmark-agnostic comparison against the best-performing PCM models from BtH, a deep neural network (DNN_PCM) and a Random Forest (RF_PCM). RandomNets with two layers gave the best overall performance, notably improving on the Matthews correlation coefficient (MCC) of the DNN baseline for the temporal split by 5.55% on the original dataset and 8.73% on the new dataset, while taking 2× less time to train on the original dataset and the same time on the new dataset. These results affirm the potential of RandomNets and extend it to PCM modelling.
ER -