1 of 8

Evo-Driven Protein Variant Generation and Structure Validation

2 of 8

Team Members

Kashvi Shah

Nikhil

Ethan Yung

Sahasra Bobbala

Arnav Kumar

Senior majoring in Computer Engineering

Minoring in Math

Sophomore in Electrical Engineering

Minoring in CS

Rising Sophomore in Electrical Engineering

Minoring in Math

Rising Sophomore in Computer Engineering

Minoring in CS

Rising Senior Majoring in Computer Science

3 of 8

The Problem

How Students Learn Today

A Bio-101 student studying cellular respiration looks at static diagrams in a textbook, watches a Khan Academy video, or stares at a flat screen.

�

When only one student has a VR headset, interactive 3D protein visualization exists — but it's inaccessible to the other 499 in the lecture hall.

�

Root cause: not enough GPU power to serve a full class simultaneously.

What This Project Enables

Imagine a live Bio-101 lecture:

the professor pulls up a digital twin of a cell. Every student — on laptop or AR/VR headset — gets their own interactive simulation.

�

Students adjust oxygen levels with a slider and watch glucose → ATP shut down in real time. They see the cell switch to anaerobic pathways — live, in 3D.

�

500 students. 500 simultaneous simulations. All powered by the H100 GPU cluster.

→

4 of 8

Project Description

  • Select a target protein group from the Protein Data Bank
  • Generate new sequence variants using Arc Evo 2
  • Predict their 3D structures with AlphaFold3 / ESMFold
  • Validate predictions against real structures from the Rutgers PDB

​

5 of 8

Project Architecture

no

yes

no

Microservice Architecture: H100 servers run inference for Evo2, AlphaFold3, and ESMFold models.

​

Expose API endpoints at these servers.

​

Visualize the generated and folded proteins in AR/VR. This requires building worlds in Unity and deploying to HoloLens/Oculus.

​

Evaluate structure for structural and biological plausibility using metrics from protein generation and folding models.

no

1.

2.

3.

4.

6 of 8

What we did

  • Installed the needed softwares
  • Configured SSH Keys
  • Set up our Wiki
  • Set up GitHub

7 of 8

Goals for next week

  • Pull and explore Protein Data Bank data on UniProt, RCSB PDB
  • Learning the basics of the GPUs
  • Pick a protein family and start working on the database

8 of 8

Thank You

Questions?

Introduction to DNA

8

Questions?