• HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
Sunday, November 22, 2026
BIOENGINEER.ORG
No Result
View All Result
  • Login
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
No Result
View All Result
Bioengineer.org
No Result
View All Result
Home NEWS Science News Chemistry

Machine Learning Model Predicts Monoamine Transporter Activity and Target Selectivity

by
November 22, 2026
in Chemistry
Reading Time: 2 mins read
0
Machine Learning Model Predicts Monoamine Transporter Activity and Target Selectivity

Machine Learning Model Predicts Monoamine Transporter Activity and Target Selectivity

Share on FacebookShare on TwitterShare on LinkedinShare on RedditShare on Telegram

Researchers have developed DepFix, a multitask machine learning framework designed to jointly predict chemical compound activity across the serotonin, dopamine, and norepinephrine transporters while deriving relative pharmacological selectivity. Monoamine transporters represent well-established therapeutic targets in antidepressant pharmacology, but publicly accessible bioactivity data across these three proteins remain notably sparse and unevenly distributed. By modeling activity differences simultaneously across the target family, the framework addresses data limitations to assist researchers in prioritizing compounds with distinct pharmacological profiles for early antidepressant discovery.

The computational architecture combines a multitask directed message-passing neural network with target-specific Morgan fingerprint random forest models. Trained on a curated dataset of 6,679 chemical compounds, DepFix achieved scaffold-test macro-averaged R² values of 0.560 for overall activity and 0.553 for activity-derived selectivity. Methodological comparisons demonstrated that blending these two distinct modeling branches delivered the clearest predictive improvement, whereas incorporating gated task adapters and an auxiliary loss yielded smaller performance gains under random splitting benchmark conditions.

Model robustness testing demonstrated that DepFix’s true-label predictive performance exceeded results from all 100 Y-randomization benchmark tests, confirming that its predictions reflect genuine chemical relationships rather than statistical artifacts. In additional analyses, temporal validation and functional-assay evaluations indicated that model transfer was endpoint-dependent across experimental contexts. To test the approach on large chemical libraries, the investigators screened 24,385 unique structures from the HERB database, effectively prioritizing candidate molecules characterized by serotonin-preferential and triple-balanced activity patterns.

Structural modeling incorporating molecular docking and molecular dynamics simulations further characterized plausible ligand poses inside the transporter binding pockets. To examine biological responses, the authors investigated an exploratory four-compound panel in retinoic-acid-differentiated SH-SY5Y cell cultures. In these cell-based assays, pterostilbene induced predominantly serotonergic alterations, whereas dihydro-N-caffeoyltyramine and 3,5-dihydroxy-1-(3,4-dihydroxyphenyl)-7-(4-hydroxyphenyl)heptane modulated all three monoamine systems, providing empirical support for the computationally predicted multi-target activity profiles.

Collectively, the study demonstrates that DepFix can jointly predict binding activity and relative selectivity across multiple monoamine transporters despite sparse training data. Although temporal and functional transfer analyses highlight important endpoint-dependent limitations in generalizability across distinct assay settings, the machine learning framework provides valuable computational decision support for screening and prioritizing promising multi-target small molecules during early antidepressant development.

Subject of Research: Machine learning prediction of monoamine transporter activity and selectivity

Article Title: DepFix: a multitask machine learning framework for joint prediction of monoamine transporter activity and selectivity

Article References: Liang, H., Ma, J., Li, R., Feng, J., & Liu, Y. (2026). DepFix: a multitask machine learning framework for joint prediction of monoamine transporter activity and selectivity. Molecular Diversity, 31(1), Article 1. https://doi.org/10.1007/s11030-026-11747-4

Image Credits: AI Generated

DOI: 10.1007/s11030-026-11747-4

Keywords: DepFix, machine learning, monoamine transporters, antidepressant pharmacology, serotonin transporter, dopamine transporter, norepinephrine transporter, directed message-passing neural network, Morgan fingerprints, random forests, scaffold test, HERB database, molecular dynamics simulations, molecular docking, pterostilbene, SH-SY5Y cells, multitask learning

News Source: Blake Davidson. (November 22, 2026). Machine Learning Model Predicts Monoamine Transporter Activity and Target Selectivity. Scienmag.

Tags: antidepressant pharmacologyDepFixdirected message-passing neural networkdopamine transporterHERB databaseMachine Learningmolecular dockingmolecular dynamics simulationsmonoamine transportersMorgan fingerprintsmultitask learningnorepinephrine transporterpterostilbenerandom forestsscaffold testserotonin transporterSH-SY5Y cells
Share12Tweet7Share2ShareShareShare1

Related Posts

Machine Learning Screening Identifies Novel Epigenetic Candidates Targeting Cancer Enzyme HDAC1

Machine Learning Screening Identifies Novel Epigenetic Candidates Targeting Cancer Enzyme HDAC1

November 22, 2026
Phloem Girdling Drives Rhizosphere Microbial Adaptation to Belowground Carbon Limitation in Tobacco

Phloem Girdling Drives Rhizosphere Microbial Adaptation to Belowground Carbon Limitation in Tobacco

November 22, 2026

Exfoliated Dual-Ligand Cobalt and Nickel Nanosheets Boost Photoelectrocatalytic Water Oxidation

November 22, 2026

Rosemary-Derived Cobalt Complexes Enhance Crystallinity and Optical Response in Biopolymers

November 22, 2026

POPULAR NEWS

  • Loneliness Strongly Associated With Depression Among Older Adults in Golestan Iran

    Loneliness Strongly Associated With Depression Among Older Adults in Golestan Iran

    29 shares
    Share 12 Tweet 7
  • Self-Powered Neuroprosthesis Harvests Walking Energy to Improve Mobility After Stroke

    29 shares
    Share 12 Tweet 7
  • Comparative Genomics Unveils Evolutionary Lineages and Stress Genes Across Four Vigna Legumes

    29 shares
    Share 12 Tweet 7
  • Trifluoroacetate Counterion Reduces Atherosclerosis and Plasma Lipids in Mice via PPAR-α

    29 shares
    Share 12 Tweet 7

About

We bring you the latest biotechnology news from best research centers and universities around the world. Check our website.

Follow us

Recent News

Truncated Pyramid Voids Preserve Flexural Strength in Concrete Slabs Under One-Way Bending

Microcurrent Signals Precede Failure in Loaded Composite Coal-Rock Structures

Machine Learning Screening Identifies Novel Epigenetic Candidates Targeting Cancer Enzyme HDAC1

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 89 other subscribers
  • Contact Us

Bioengineer.org © Copyright 2023 All Rights Reserved.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • Homepages
    • Home Page 1
    • Home Page 2
  • News
  • National
  • Business
  • Health
  • Lifestyle
  • Science

Bioengineer.org © Copyright 2023 All Rights Reserved.