PhD Studentship: AI for Drug and Vaccine Discovery: Novel Targets to Control Coccidiosis in Chickens

Coccidiosis caused by Eimeria is the most important parasitic disease in chickens, costing the UK poultry industry ~£100 million annually (Blake et al., 2020). Recent figures indicate that 97% of UK broilers are reared using anticoccidial drugs. However, reliance on anticoccidial drugs is increasingly problematic: resistance is common and products available for use are dwindling. Alternatives include vaccination using live parasite vaccines, although uptake has been limited. New options are required for control of coccidiosis, including new drugs and vaccines that are cheaper and more easily scalable. Data-led approaches to identify and prioritise drug and vaccine targets are becoming available, offering new opportunities.

Protein-protein interactions (PPI) underlie most cellular functions where host-pathogen interactions commonly determine the outcomes of infection. This project will focus on host-pathogen interactions during Eimeria infection to identify and prioritise hubs and bottleneck proteins in the network as new drug and vaccine targets. Graph-based machine learning (Dong et al., 2020, Atz et al., 2021) allows inference of structure and exploitation for rewiring events during infection (Cuesta-Astroz et al., 2019). We have started to generate host-pathogen interactome maps for Eimeria tenella infection in chickens using graph-based machine learning, integrating network topology analysis with protein function-based node embeddings. The models generated advance our ability to identify key proteins that act as ‘brokers’ in essential host and parasite PPIs. A complementary bioinformatic vaccinology approach (e.g., Goodswen et al., 2023) will be used to prioritise candidates for vaccine development. The studentship will benefit from a BBSRC funded project developing Saccharomyces cerevisiae as a vaccine delivery vector, using the vector as a convenient ‘plug & play’ approach to screen candidates for vaccine efficacy. 

HYPOTHESIS

Graph-based machine learning can be trained to use existing and new host-pathogen transcriptomic datasets to identify key proteins that feature in Eimeria PPIs as candidates for drug and vaccine development. 

OBJECTIVES

This proposal aims to build on our development of graph-based machine learning, using RNAseq transcriptomic datasets to investigate rewiring events of PPI networks and identify key Eimeria proteins that contribute to essential pathways as candidates for drug targets. Novel Eimeria antigens identified at the interface of host-pathogen PPI networks will be screened using a vaccine ranking pipeline.

1. Produce new RNAseq transcriptome datasets for training and testing of rewiring events in host-pathogen networks.
2. Optimise existing graph-based machine learning algorithm to phenotypically model interactome maps for Eimeria species parasites during infection, using E. tenella as an example.
3. Screen new improved Eimeria genome assemblies and annotations using a bioinformatic pipeline with custom modifications to identify and prioritise drug and vaccine candidates.
4. Combine #2 and #3 to prioritise protein candidates for screening as novel drugs targets (working with the iCASE partner, Boehringer Ingelheim Animal Health) and/or vaccine candidates.
5. Clone candidate coding sequences into the S. cerevisiae vaccine vector, confirming expression by FACS and Western blotting.
6. Undertake a pilot in vivo screen of vaccine efficacy.

Advert information