Chapter 35 CWD Sentinel

IN WAREHOUSE DEVELOPMENT

35.1 What is the CWD Sentinel?

The CWD Sentinel uses deep learning with INLA to predict which sub-admin areas may turn positive for the first time, or if already positive, may have infections that spread quickly.

35.2 What Questions Does it Answer?

Question 1. Where is CWD likely to emerge given historical disease emergence around the region? Predictions emphasize areas that warrant targeted CWD surveillance because of similar conditions with areas known to harbor CWD.

Question 2. Where is CWD likely to intensify? Predictions emphasize areas that warrant targeted CWD surveillance because of similar conditions with areas known to harbor increasing CWD.

35.3 Output Details

A map containing the predictions of CWD emergence and spread for each sub-administrative area in the study area.

35.4 Abbreviated Tutorial

  1. Run the CWD Sentinel from the CWD Data Warehouse.
  2. Explore the model logs, input file, and output files used in the run.
  3. If the model did not run, check the model logs to understand required data that was missing.
  4. Explore the machine learning predictions of disease status.

35.5 Parameters Needed to Execute the Model

  • Model type: Select ‘CWD Sentinel’ from the drop-down list.

  • Reference name: Label the run.

  • (Optional) Applicable season-year: Label the season-year of the run to assist in documentation.

  • (Optional) Notes: Enter any remarks about the run.

  • Season-year: Select the season-year to be used in the estimation.

35.6 Details on the Theory

Gonzalez-Crespo C, Schuler K, Hanley B, Hollingshead N, Middaugh C, Ballard J, Clemons B, Kelly J, Harms T, Caudell J, Benavidez Westrich K, McCallen E, Casey C, O’Brien L, Trudeau J, Stewart C, Carstensen M, Jennelle C, McKinley W, Hynes P, Stevens A, Miller L, Grove D, Storm D, Martinez-Lopez B. Fusing Bayesian inference and deep learning: A hybrid AI approach for predicting chronic wasting disease emergence and spread. In revision