Nat Commun May 2023 PMID 37188731

A general model to predict small molecule substrates of enzymes based on machine and deep learning.

Alexander Kroll, Sahasra Ranjan, Martin K M Engqvist, Martin J Lercher

Full text PDF Download social card

Mentions

Peak: 20 in May 2023

across 5 sources since May 2023

Extended reach

32,615

potential audience

Total likes

Video views

Mentions per month

May 2023 – Apr 2024

Where it's discussed

X / Twitter22 posts
News16 articles
Reddit2 posts
Blogs2 articles
YouTube1 videos

AI synopsis

What people are saying

Most discussions revolve around sharing and briefly summarizing the research, with phrases like 'new machine learning model' and 'predict enzyme functions,' demonstrating a moderate level of curiosity and acknowledgment of its significance. The tone includes technical references such as 'transformer-based approach' and mentions of platforms like 'Web,' indicating the perceived importance of the model, though they don't seem to delve deeply into its implications or potential controversies.

AgreementNeither agree nor disagree

The majority of discussions are neutral or slightly positive, with some expressions of support for the research but no strong endorsements or criticisms.

InterestModerate interest

Posts show moderate curiosity, with many sharing the publication and mentioning keywords like 'machine learning' and 'enzyme functions,' indicating genuine interest.

EngagementModerate engagement

Several discussions include retweets and comments referencing concepts such as transformer models and practical applications, suggesting some active engagement.

ImpactModerate impact

The discussions highlight the potential significance of the model, especially given mentions of its accuracy and broad applicability, but lack evidence of deep influence or real-world adoption at this stage.

Abstract in brief

The article introduces ESP, a machine learning model designed to predict which small molecules are substrates for enzymes, especially when experimental data is limited or costly to obtain.
ESP uses a transformer-based approach and is trained on both confirmed substrates and randomly sampled molecules to better distinguish true substrates from non-substrates, achieving over 91% accuracy.
The model works across various enzymes and metabolite types, and a web server version allows researchers to perform quick, in silico substrate predictions to support scientific and practical applications.]

Read the full abstract

All mentions

Machine Learning Model for Predicting Enzyme Small Molecule Substrates

The article introduces ESP, a machine learning model designed to predict which small molecules are substrates for enzymes, especially when experimental data is limited or costly to obtain. ESP uses a transformer-based…

STEM RTCL TV · Aug 21, 2023 26 views

John Maki

@John_W_Maki · X

A general model to predict small molecule substrates of enzymes based on machine and deep learning | Nature Communications https://t.co/lbdV1UNecc

Apr 6, 2024 · 238 followers View post →

kwh_rd100

@KwhRd100 · X

Jun 6, 2023 · 2 likes · 286 followers View post →

Ilnam Kang

@ilnam_kang · X

RT @pacyc1841: » A general model to predict small molecule substrates of enzymes based on machine and deep learning | Nature Communications…

May 24, 2023 · 1 share · 235 followers View post →

Tominaga K. (tomiken)

@pacyc1841 · X

» A general model to predict small molecule substrates of enzymes based on machine and deep learning | Nature Communications https://t.co/7DxMzVrI8S

May 24, 2023 · 1 share · 855 followers View post →

Natural Resource Today

@ResourceToday · X

Researchers have developed a new machine learning model to predict #enzyme functions. The @HHU_de-led research is published in the journal @NatureComms. https://t.co/ajuSiWhm7V

May 23, 2023 · 1 like · 520 followers View post →

Ordo Fraterna Fibonacci

@OrdoFibonacci · X

A general model to predict small molecule substrates of enzymes based on machine and deep learning https://t.co/Ko8v8F0sxh

May 23, 2023 · 2,569 followers View post →

Marc Dämgen

@DaemgenMarc · X