Deep learning-based classification of peptide analytes from single-channel nanopore translocation events

What is it about?

We've trained an AI to act like a detective, using a microscopic sensor to identify tiny protein fragments as they pass through a hole one by one. The AI learns to recognize the unique electrical "fingerprint" that each different molecule creates during its journey. While identifying a single fragment in a complex mixture can be challenging, the AI becomes perfectly accurate when it analyzes a stream of signals and takes a majority vote. This is a key step towards building powerful new tools for diagnosing diseases by rapidly detecting specific protein biomarkers.

Featured Image

Photo by Niloy T on Unsplash

Photo by Niloy T on Unsplash

Why is it important?

Towards building powerful new high-throughput, low-cost tools for diagnosing diseases by rapidly detecting specific protein biomarkers, we describe a computational pipeline to detect peptides from single translocation events through nanopores.