Search

MS-Net: Multi-Similarity-Based Network Annotation for Untargeted Metabolomics.

Pereira Francisco V, Duthen S, Crossay E, Perez A, Hennechart S

Metabolomics

Every herb, vegetable, and wildflower in your garden contains hundreds of unknown compounds that may influence flavor, fragrance, or medicinal effect — tools like this are what finally let researchers decode that hidden chemistry plant by plant.

When scientists grind up a plant and analyze its chemistry, they can detect thousands of molecules but struggle to figure out what most of them actually are. MS-Net is a new computer workflow that cross-references multiple clues — how molecules are shaped, how they behave in the instrument, and what's already known about that plant family — to make better guesses. Tested on cannabis, it correctly identified over 1,200 compounds and rescued more than half of the right answers from deep in the pile of wrong guesses.

Key Findings

1

MS-Net identified 1,275 compounds from a starting pool of over 118,000 candidates in cannabis extracts, collapsing 2,595 detected features to 1,297 after filtering.

2

53% of correct final annotations were 'rescued' from ranks 2–50 in the initial prediction list, showing the tool fixes systematic errors in existing methods.

3

The workflow successfully reconstructed known cannabinoid biosynthetic pathways, confirming the identified compounds fit real plant biology.

chevron_right Technical Summary

Scientists built MS-Net, a free software tool that dramatically improves how researchers identify the thousands of chemical compounds in a plant extract. By combining multiple types of similarity data and biological knowledge, it correctly identified far more compounds in cannabis than previous methods alone.

description

Abstract Preview

Confident metabolite annotation remains a critical bottleneck in untargeted LC-MS metabolomics, with experimental spectral libraries covering only 5-20% of detected features. While in silico tools ...

open_in_new Read full abstract

Abstract copyright held by the original publisher.

hub This connects to 12 other discoveries — Cannabis, Hemp metabolomics, medicinal-plants, plant-chemistry +2 more 5 related articles

Species Mentioned

Was this useful?

mail Weekly plant science — one email, Saturdays.

Share: X/Twitter Reddit
arrow_forward Next Discovery

Chloroplast Genome Editing Eliminates Gluten Immunogenicity in Triticum aestivum

It could mean that people with celiac disease — roughly 1 in 100 worldwide — may one day safely eat bread made from real wheat, without sacrificing the taste...

eco Cannabis
Species
Cannabis

Cannabis is a genus of flowering plants in the family Cannabaceae that is widely accepted as being indigenous to and originating from the continent of Asia. However, the number of species is disputed, with as many as three species being recognized: Cannabis sativa, C. indica, and C. ruderalis. Al...