SIGMAP: An AI Tool for Predicting Sigma-1 Receptor Affinity
Author Information
Author(s): Lomuscio Maria Cristina, Corriero Nicola, Nanna Vittoria, Piccinno Antonio, Saviano Michele, Lanzilotti Rosa, Abate Carmen, Alberga Domenico, Mangiatordi Giuseppe Felice
Primary Institution: Università degli Studi di Bari Aldo Moro
Hypothesis
Can machine learning models accurately predict sigma-1 receptor affinity using a curated dataset?
Conclusion
The study developed a user-friendly AI tool, SIGMAP, that effectively predicts sigma-1 receptor affinity with high accuracy.
Supporting Evidence
- The best-performing model achieved an AUC of 0.90.
- Two explainable AI methods were used to clarify predictions.
- SIGMAP is the first freely accessible tool for predicting S1R affinity.
Takeaway
Researchers created a smart computer program that helps scientists find new medicines by predicting how well they can bind to a specific brain receptor.
Methodology
The study used 25 machine learning classifiers trained on a dataset of 2018 compounds to predict sigma-1 receptor affinity.
Potential Biases
Potential bias in the dataset due to the selection criteria for compounds.
Limitations
The models may not perform well on compounds outside the applicability domain.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website