A national early-warning model for Nigeria built on real, cross-validated NCDC weekly surveillance data (AUROC 0.88 vs 0.85 naive), with an illustrative per-state view and Lassa antiviral drug discovery.
National model: will the next 4 weeks carry an above-median confirmed-case burden (an "elevated-transmission week")? Trained on real NCDC weekly data with a strict out-of-time hold-out (train ≤2023, test 2024) and benchmarked against a naive persistence baseline.
Honest read. On real, cross-validated NCDC data the model reaches AUROC 0.880 on a 2024 out-of-time test — but a naive baseline (recent 4-week case average) reaches 0.849, so the model adds only ~0.03. On the 52-week hold-out that difference is not statistically significant (DeLong paired test p = 0.61; bootstrap 95% CI for the gap −0.08 to +0.15). Much of the signal is seasonality (the dry-season flag is the top predictor) and recent incidence. This is a modest, believable result, not a breakthrough; genuine forecasting value can only be shown prospectively, which is why a prospective-validation protocol is planned, with append-only forecast logging already in place to support it.
Open-access preprint. The full method, results, and honest limitations are published open access: Forecasting Elevated Lassa Fever Transmission Weeks in Nigeria from National Surveillance Data: A Baseline-Benchmarked Proof of Concept (Zenodo, CC-BY, DOI 10.5281/zenodo.21122486). Code & data: github.com/oluwafemidiakhoa/lassaai.
Data: NCDC Weekly Epidemiological Reports (2020–2025), cross-validated against an independent extraction and NCDC published annual totals (agreement ±~1%), plus SORMAS individual-level data (2018–2021) for a per-state/clinical cross-check. Top predictor: dry-season timing (≈53%).
Limitations: Validated result is national; the per-state map below is illustrative pending real per-state weekly data (NCDC reports it in image tables). Confirmed counts under-report true incidence. The model beats a naive baseline only slightly. Screening/research aid only — not a validated diagnostic or a replacement for NCDC field surveillance.
Enter LASV protein targets into GaiaLab's AI analysis pipeline. The viral pathogen mode routes GPC, NP, L, and Z through a Lassa-specific drug repurposing engine with known antiviral candidates pre-loaded.