Hello everyone,

I'm currently preparing a PhD project proposal and would really appreciate feedback on potential directions from the bioinformatics community.

Context:

My Master's research focused on antimicrobial resistance (AMR) genes in MRSA strains. For my PhD, I'm interested in exploring horizontal gene transfer and AMR dissemination between wildlife and livestock via the gut microbiome.

I have access to multiple shotgun metagenomic datasets from fecal samples (from various animal species, including wild and domestic), but environmental metadata (e.g., location, diet, contact info) is either sparse or missing. The challenge is to find a computational modeling angle that is still scientifically valuable and publishable — ideally one that doesn't rely heavily on metadata.

Goal:

To formulate a unique bioinformatics-driven hypothesis or framework for studying the transmission of resistance genes or strains between different animal hosts — perhaps even inferring directionality or host association of mobile genetic elements.

What I've considered so far:

AMR gene profiling using tools like ResFinder, RGI, or CARD.

Metagenomic strain-level profiling using StrainPhlAn, PanPhlAn, or inStrain to detect shared strains.

SourceTracker-like modeling to infer microbial exchange, but this might require metadata I lack.

Graph/network-based models of ARG sharing between host groups.

Potentially modeling horizontal gene transfer via mobile genetic elements (plasmids, phages) — though that seems technically intensive.

What I need help with:

Are any of the above ideas novel or interesting enough for a PhD proposal, especially given the lack of metadata?

Can strain-level or mobile element-based analysis be a strong enough angle on its own?

Do you know of any less-explored or creative frameworks that could uncover AMR dissemination routes between species without relying on geographic metadata?

Any input, related papers, or even high-level advice would be incredibly appreciated!

Thank you in advance