PhytoGenix

• 

  Dashboard of PhytoGenix

Inspiration

Crop diseases cause over $220 billion in annual losses and traditional breeding takes 7–12 years of physical crossing and field testing, so we built a tool to predict resistance and simulate crosses digitally instead.

What it does

PhytoGenix scores disease resistance from genomic markers, catalogs R-genes and pathogens, simulates virtual crosses across generations, optimizes for both immunity and yield, visualizes chromosomes interactively, models outbreak survival under different conditions, and lets breeders import/export their own genotype data locally.

How we built it

We engineered a JavaScript genetics engine with additive and epistatic scoring models plus Mendelian cross simulation, backed it with a curated genomic knowledge base spanning 8 crops and 20 pathogens, wrapped it in a Chart.js-and-Canvas-powered UI with a botanical daylight theme, and served it all through a lightweight Node.js backend.

Challenges we ran into

The hardest parts were making dense molecular genetics approachable without losing precision, building a reliable Pareto-frontier solver for yield-vs-resistance trade-offs, scaling chromosome visualizations across wildly different ploidy levels, and keeping all genotype data processing strictly client-side for privacy.

Accomplishments that we're proud of

We're proud of hitting near-instant simulation speeds for 200 offspring, achieving a clean and consistent daylight-mode aesthetic, delivering a complete breeder workflow in one app, and accurately mapping all 20 crop-pathogen systems from the source dataset.

What we learned

We came away understanding the trade-off between race-specific and broad-spectrum resistance genes, how quantitative genetics math can compress years of breeding into seconds, and why Pareto optimization matters for avoiding resistant-but-low-yield selections.

What's next for PhytoGenix

Next up is adding genetic linkage mapping, direct high-density VCF ingestion from sequencing platforms, live satellite weather integration for outbreak modeling, and an AI copilot for conversational breeding strategy.

Built With

• csv-vcf-parsing
• epistasis-modeling
• firebase-hosting
• gebv-modeling
• genomic-selection
• google-cloud-run
• html5
• javascript
• mendelian-simulation
• node.js
• pareto-optimization
• rest-api
• serverless-vercel-lambda