AI-Assisted Crystal Detection in the TEM to Accelerate MicroED

Image of a structure of a protein determined by MicroED.

Leverage the power of computer vision and deep learning to automate the discovery of protein crystals in electron microscopes, accelerating the pace of modern drug discovery. 

Project is Not Currently Available

This project has reached full capacity for the current term. Please check back next semester for updates.

Project description

Microcrystal Electron Diffraction (MicroED) is a cutting-edge technique that allows scientists to determine the atomic structures of proteins and small molecules using standard Transmission Electron Microscopes (TEM). While this method has the potential to revolutionize drug discovery and structural biology, a major bottleneck remains: the manual identification of usable nanocrystals on a sample grid. Currently, researchers must spend hours manually scanning grids to find crystals suitable for diffraction.

This project seeks to overcome this limitation by developing an AI-assisted computer vision pipeline to automate crystal detection. The student will utilize Python and deep learning frameworks (such as YOLO or TensorFlow) to train an object detection model using a dataset of electron micrographs. The goal is to create a tool that can instantly recognize and bound protein crystals in low-dose TEM images. By automating the "eyes" of the microscope, this project will drastically increase the throughput of structure determination, allowing for faster analysis of therapeutic targets and biological mechanisms. 

Project outcome

A trained and validated neural network model capable of identifying crystals in TEM images with high accuracy.

A documented Python codebase/notebook demonstrating the inference pipeline.

A research poster presentation at the ELN Celebration of Excellence.

Potential co-authorship on a manuscript detailing the automation of MicroED workflows. 

Project details

Timing, eligibility and other details
Length of commitment About a semester; 3-5 months
Start time Anytime
In-person, remote, or hybrid? Hybrid
Level of collaboration Individual student project
Benefits Work Study or Stipend
Who is eligible All undergraduate students 

Core partners

  • The Martynowycz Lab (Department of Structural Biology, Jacobs School of Medicine & Biomedical Sciences)

Project mentor

Michael Martynowycz

Assistant Professor

Structural Biology

Phone: (630) 415-8495

Email: mmartyno@buffalo.edu

Start the project

  1. Email the project mentor using the contact information above to express your interest and get approval to work on the project. (Here are helpful tips on how to contact a project mentor.)
  2. After you receive approval from the mentor to start this project, click the button to start the digital badge. (Learn more about the digital badge.) 
Start the Digital Badge

Preparation activities

Once you begin the digital badge series, you will have access to all the necessary activities and instructions. Your mentor has indicated they would like you to also complete the specific preparation activities below. After you’re approved to begin the project, your mentor will send the relevant materials. Please reference this when you get to Step 2 of the Preparation Phase. 

  • Attending workshops and trainings
    • Completing the necessary lab safety or computer safety courses required by UB for facility access to HWI. 

  • Familiarizing yourself with the concepts of "Object Detection" in computer vision, specifically the YOLO (You Only Look Once) architecture.

  • Downloading and installing Anaconda (or Miniconda) on your personal computer or verify access to UB’s CCR computing resources.

  • Creating a generic Python environment and successfully launching a Jupyter Notebook. Installing the OpenCV-Python and Matplotlib libraries and verifying they import without errors

Keywords

AI, computer vision, computation, structural biology, object detection, code