This research paper is the final chapter of my PhD thesis which, excitingly, will soon be out of its embargo! Thank you to all co-authors and funding from NERC and Cefas!

Alongside well-reported drivers of V. vulnificus (sea temperature), we also found salinity, chlorophyll-a concentration and coastal runoff to be strong drivers within the model. This accentuates potential of robust surveillance for forecasting environmentally-associated marine infections.

The highest-performing model had an AUC score of 0.984 and a sensitivity of 0.971, highlighting potential of machine learning to anticipate areas/periods of V. vulnificus risk. Different spatial scales and model set-ups were tested, as well as potential of early warning systems using lagged data.

We trained machine learning models (random forest classifiers) with this data, 6 environmental variables (sea temperature, salinity, chlorophyll a concentration, sea level anomalies, land temperature and runoff) and categorical encoders to assess our potential to forecast V. vulnificus infections.

While vibriosis infections are usually rarely reported, the COVIS database has reported V. vulnificus infections in the USA since 1988, offering a unique opportunity.

Published in @jmirpub.bsky.social Bioinformatics and Biotechnology yesterday, in theme issue “Machine Learning-Based Predictive Models using Genomic Data”. Thanks to co-authors @jaimemurtaza.bsky.social @aquagenomics.bsky.social and Chris Hauton 👏

Or read our plain English 'Behind the Paper' blog post here: communities.springernature.com/posts/redefi...

Our study redefines the story of VpST3 expansion in Latin America, and provides insight into the ecological and evolutionary dynamics that allow a marine pathogen to emerge in regions with distinct climates. 🧬🌊💡 Access the full paper PDF here (fully open access): rdcu.be/dThip

While this population was so specialised to LatAm, in recent years, a newer strain has emerged which is becoming dominant both globally and in LatAm- the VpST3 story continues… 📖

Signatures of selection to Latin America's distinct marine environment (e.g. colder temperature, higher salinity, lower pH) were found in mutations and accessory genes in this population, providing adaptive benefits in terms of survival or transport 💪

We found a genetically distinct population that emerged and became dominant in Latin America 📈 Unlike previous consensus, this population did not evolve from initial strains found in India, but from a separate ancestral variant in LatAm- diverging from the Asian populations ❗️

The ST3 strain was assumed to emerge in India in 1996, but was reported in Peru (thousands of miles away) in the same year 🤔 We used novel genomic samples from Latin America and environmental data to reconstruct the global expansion dynamics of VpST3 🧬 🌡️

Vibrio parahaemolyticus is a unique water-borne human pathogen that is capable of pandemic expansion around the globe 🌊🚢 🌎 This bacteria lives in coastal waters, so is driven by changes in its marine environment 🌡️