Improving metagenomic taxonomic annotations with deep learning

Improving metagenomic taxonomic annotations with deep learning

Spread the love



This is a 5 minutes explainer of the technical details of our latest published article: Kutuzova et al. “Taxometer: Improving taxonomic annotations of metagenome contigs”, Nature Communications. https://www.nature.com/articles/s41467-024-52771-y

More about hierarchical losses: https://youtu.be/NAiw8atLktk?si=KqdGIMyKxBHkRGGg

The software source code is at https://github.com/RasmussenLab/vamb

The animations are made with a fantastic Manim library https://www.manim.community/. All the code is here https://github.com/sgalkina/animations/blob/main/notebooks/Taxometer.ipynb

00:00 – Features: abundance vector
00:41 – Features: tetranucleotide frequency
01:05 – Neural network
01:45 – Hierarchical loss
03:26 – Workflow

Source

Comments

No comments yet. Why don’t you start the discussion?

Leave a Reply

Your email address will not be published. Required fields are marked *