Putting computational models of immunity to the test—An invited challenge to predict B.pertussis vaccination responses DOI Creative Commons
Pramod Shinde, Lisa Willemsen, Michael C. Anderson

et al.

PLoS Computational Biology, Journal Year: 2025, Volume and Issue: 21(3), P. e1012927 - e1012927

Published: March 31, 2025

Systems vaccinology studies have been used to build computational models that predict individual vaccine responses and identify the factors contributing differences in outcome. Comparing such is challenging due variability study designs. To address this, we established a community resource compare predicting B. pertussis booster generate experimental data for explicit purpose of model evaluation. We here describe our second prediction challenge using this resource, where benchmarked 49 algorithms from 53 scientists. found most successful stood out their handling nonlinearities, reducing large feature sets representative subsets, advanced preprocessing. In contrast, adopted literature were developed antibody other settings performed poorly, reinforcing need purpose-built models. Overall, demonstrates value purpose-generated datasets rigorous open evaluations features improve reliability applicability response prediction.

Language: Английский

Putting computational models of immunity to the test—An invited challenge to predict B.pertussis vaccination responses DOI Creative Commons
Pramod Shinde, Lisa Willemsen, Michael C. Anderson

et al.

PLoS Computational Biology, Journal Year: 2025, Volume and Issue: 21(3), P. e1012927 - e1012927

Published: March 31, 2025

Systems vaccinology studies have been used to build computational models that predict individual vaccine responses and identify the factors contributing differences in outcome. Comparing such is challenging due variability study designs. To address this, we established a community resource compare predicting B. pertussis booster generate experimental data for explicit purpose of model evaluation. We here describe our second prediction challenge using this resource, where benchmarked 49 algorithms from 53 scientists. found most successful stood out their handling nonlinearities, reducing large feature sets representative subsets, advanced preprocessing. In contrast, adopted literature were developed antibody other settings performed poorly, reinforcing need purpose-built models. Overall, demonstrates value purpose-generated datasets rigorous open evaluations features improve reliability applicability response prediction.

Language: Английский

Citations

0