Skip to content

deep-learning for metadata extraction from the C. elegans embryo

Notifications You must be signed in to change notification settings

Mainakdeb/elegant-embryos

Repository files navigation

elegant-embryos

This repository contains work surrounding DevoLearn models and metadata extraction from the C. elegans embryo.

Binder

The EPIC dataset contains time series data surrounding the embryogenesis of the C. elegans embryo. I had to Extract the X, Y, Z coordinates of each cell (of all lineages), extract the timestamps of each cell in sequence, using which I generated the following plots.

  1. A plot showcasing cell positions through time.

  1. Visualizing single cell movement through time. This example shows the movement of the cell of lineage 'Dpa' through time.

Gaining insights from static plots can be difficult at times, this is where these interactive plots can help by enabling users to rotate, zoom, animate plots (with respect to time), rather than being stuck with just one point of time and perspective.

Post processing algorithms modify the outputs from the neural-network in order to further increase the accuracy of the predictions. I implemented a post processing algorithm on the lineage population model from DevoLearn, which exploits an inherent pattern in the data and modifies the model outputs accordingly, resulting in better accuracy (about 30% less loss) as highlighted in figure below.

About

deep-learning for metadata extraction from the C. elegans embryo

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published