BrainRoute integrates machine learning prediction with community-driven verification of blood-brain barrier permeability

The blood-brain barrier (BBB) is a highly selective barrier that restricts the entry of many therapeutic agents into the brain. The lack of efficient methods for BBB permeability research and data management remains a significant challenge for central nervous system drug-development pipelines. Although BBB permeability data are available from multiple open sources, few platforms integrate these datasets with molecular exploration, predictive tools, and a structured process for community-supported data verification. Here, we present BrainRoute, an open platform for BBB molecular data collection, exploration, prediction, and community-supported verification. BrainRoute provides searchable and filterable molecular records, physicochemical and structural information, source-linked data, single-molecule and batch prediction, and a structured workflow through which users can submit additional evidence for review. The prediction tool model-development cohort contained 7,888 unique molecules, including 5,424 BBB+ and 2,464 BBB- molecules. Six machine-learning classifiers were evaluated using PaDEL descriptors, Morgan fingerprints, frozen ChemBERTa embeddings, and combined feature representations. The three selected models were integrated into the platform as a probability-averaged ensemble. The primary five-fold scaffold-grouped analysis achieved a best mean balanced accuracy of 0.783 ± 0.026. On an independent external set containing 410 non-overlapping molecules, the deployed ensemble achieved a balanced accuracy of 0.659 (95% CI 0.600 - 0.712) and an AUPRC of 0.670 (95% CI 0.583 - 0.752). BrainRoute serves as a platform for exploring BBB-related molecular data and computational predictions while supporting the continued growth and verification of the underlying dataset. BrainRoute is freely accessible online (https://github.com/omicscodeathon/brainroutedb).

Authors

Institutions

Publication Details

Journal
bioRxiv (Cold Spring Harbor Laboratory)
Published
2026-11-20
DOI
https://doi.org/10.64898/2026.11.19.767839
Primary Topic
Computational Drug Discovery Methods
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

BrainRoute integrates machine learning prediction with community-driven verification of blood-brain barrier permeability

Olaitan Igbagbo Awe, Soham Shirolkar, Lewis Tem Bueh, Noura E. Ahmed et al.
2 links
bioRxiv (Cold Spring Harbor Laboratory)
Computational Drug Discovery Methods
preprint

BrainRoute integrates machine learning prediction with community-driven verification of blood-brain barrier permeability

Olaitan Igbagbo Awe, Soham Shirolkar, Lewis Tem Bueh, Noura E. Ahmed, Georges Some, Leah W. Cerere
preprint en

Abstract

The blood-brain barrier (BBB) is a highly selective barrier that restricts the entry of many therapeutic agents into the brain. The lack of efficient methods for BBB permeability research and data management remains a significant challenge for central nervous system drug-development pipelines. Although BBB permeability data are available from multiple open sources, few platforms integrate these datasets with molecular exploration, predictive tools, and a structured process for community-supported data verification. Here, we present BrainRoute, an open platform for BBB molecular data collection, exploration, prediction, and community-supported verification. BrainRoute provides searchable and filterable molecular records, physicochemical and structural information, source-linked data, single-molecule and batch prediction, and a structured workflow through which users can submit additional evidence for review. The prediction tool model-development cohort contained 7,888 unique molecules, including 5,424 BBB+ and 2,464 BBB- molecules. Six machine-learning classifiers were evaluated using PaDEL descriptors, Morgan fingerprints, frozen ChemBERTa embeddings, and combined feature representations. The three selected models were integrated into the platform as a probability-averaged ensemble. The primary five-fold scaffold-grouped analysis achieved a best mean balanced accuracy of 0.783 ± 0.026. On an independent external set containing 410 non-overlapping molecules, the deployed ensemble achieved a balanced accuracy of 0.659 (95% CI 0.600 - 0.712) and an AUPRC of 0.670 (95% CI 0.583 - 0.752). BrainRoute serves as a platform for exploring BBB-related molecular data and computational predictions while supporting the continued growth and verification of the underlying dataset. BrainRoute is freely accessible online (https://github.com/omicscodeathon/brainroutedb).

bioRxiv (Cold Spring Harbor Laboratory)
University of Buea (CM), Mount Kenya University (KE), University of South Florida (US), Munster Technological University (IE), Institut de Recherche en Sciences de la Santé (BF), Clinical Research Unit Nanoro (BF)
National Institutes of Health
Good health and well-being
Computational Drug Discovery Methods
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon