Sign In

Communications of the ACM

ACM TechNews

ML Helps Reveal Cells' Inner Structures in Detail

View as: Print Mobile App Share:
An image showing the interaction between mitochondria (orange) and the endoplasmic reticulum (green.)

Using online visualization tools, researchers can mine the COSEM datasets to see how different organelles interact inside a cell.

Credit: COSEM Project Team

The Howard Hughes Medical Institute (HHMI)'s Cell Organelle Segmentation in Electron Microscopy (COSEM) team combined electron microscopy with machine learning (ML) to identify inner structures within cells.

The researchers programmed new algorithms to map entire cells in hours rather than years, and opened the OpenOrganelle data portal to allow anyone access to these datasets and tools.

For 10 years, microscopes have been generating high-resolution images of fly and mammalian brains, and COSEM's Larissa Heinrich repurposed ML algorithms designed to pinpoint synapses to segment organelles.

The algorithms classify each pixel by its distance to each of 30 different kinds of organelles and structures, then integrate that data to predict organelle positions.

HHMI's Wyatt Korff said resources like CORSEM's algorithms could help solve mysteries about cellular interaction in different tissues "in a way that we haven't had access to in the past."

From Howard Hughes Medical Institute News
View Full Article


Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account