A novel strategy for classifying the output from an in silicovaccine discovery pipeline for eukaryotic pathogens using machine learning algorithms
Problem
In silico vaccine candidate pipelines for eukaryotic pathogens produce many false candidates that would require expensive laboratory validation.
Why it matters
Laboratory validation of false candidates wastes time and money in vaccine development.
What the AI did
Machine learning algorithms were trained on proteins from Toxoplasma gondii, Plasmodium sp., and C. elegans to distinguish expected true from expected false vaccine candidates.
Results
Average sensitivity 0.97 and specificity 0.98 for distinguishing expected true from false candidates.
Impact
Evidence supports ML filtering of candidates; actual vaccine discovery or lab confirmation not stated.
Record
| Venue | BMC Bioinformatics |
|---|---|
| Published | 2013 |
| Work type | article |
| Authors | Stephen J. Goodswen, Paul J Kennedy, John Ellis |
| Citations | 46 (OpenAlex, at import time) |
| References | 55 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1186/1471-2105-14-315 |
| Open access | gold |
| Keywords | machine learning, vaccine discovery, in silico pipeline, Toxoplasma gondii, candidate filtering |
Importance rating
real but narrower advance
Reducing false in silico vaccine candidates saves lab time and money in vaccine development for eukaryotic pathogens; a useful but narrow pipeline step, with no disease-burden scale stated.