Type 2 Diabetes · 17 targets · QSAR

Predict antidiabetic
activity instantly.

Paste any SMILES and get activity predictions across 17 validated T2D targets — with applicability domain charts so you know when to trust each result.

Try:
ScreenBETES

How it works

Built on rigorous QSAR methodology

01

Input SMILES

Paste any valid SMILES — natural product, synthetic compound, or drug candidate.

02

Fingerprint encoding

Your molecule is encoded into Morgan ECFP4, AtomPair or MACCS fingerprints, depending on the best model for each target.

03

QSAR prediction

Each of the 17 target models (selected by MCC across 27 algorithm × fingerprint combinations) returns a probability of activity.

04

AD histogram

The Tanimoto similarity distribution of your molecule vs. the training set is shown per target — so you see exactly how reliably each prediction can be trusted.

17 validated T2D targets

Full pathway coverage

DPP-4Dipeptidyl peptidase IV
PTP1BProtein tyrosine phosphatase 1B
SGLT1Sodium-glucose cotransporter 1
SGLT2Sodium-glucose cotransporter 2
PPARγPeroxisome proliferator-activated receptor γ
AKR1B1Aldose reductase
FFAR1Free fatty acid receptor 1
GLP-1RGlucagon-like peptide-1 receptor
MGAMMaltase-glucoamylase
mTORmTOR kinase
INSRInsulin receptor
GSK-3βGlycogen synthase kinase-3β
FXRFarnesoid X receptor
PPARαPeroxisome proliferator-activated receptor α
PPARδPeroxisome proliferator-activated receptor δ
HSD11B111β-Hydroxysteroid dehydrogenase 1
SISucrase-isomaltase

The team

Built by researchers, for researchers

CS

Carlos Seiti H. Shiraishi

QSAR Pipeline · Virtual Screening · CIMO, IPB¹

CM

Dr. Cleber C. Melo-Filho

Molecular Modeling · Chemical Biology · UNC Chapel Hill³

SH

Dr. Sandrina A. Heleno

Natural Products · Phytochemistry · CIMO, IPB¹

MP

Dr. Miguel A. Prieto

Chemoinformatics · Food Science · Universidade de Vigo²

LB

Dr. Lillian Barros

Natural Products · Bioactivity · CIMO, IPB¹

MS

Dr. Marcus T. Scotti

Cheminformatics · Natural Products · UFPB⁴

EM

Dr. Eugene N. Muratov

QSAR Methodology · Regulatory Science · UNC Chapel Hill³

RA

Dr. Rui M. V. Abreu

Pharmacology · Project Coordination · CIMO, IPB¹

¹ CIMO, LA SusTEC, Instituto Politécnico de Bragança, 5300-253 Bragança, Portugal

² Nutrition and Bromatology Group, Universidade de Vigo, 36310 Vigo, Spain

³ UNC Eshelman School of Pharmacy, University of North Carolina, Chapel Hill, NC 27599, USA

⁴ Laboratório de Quimioinformática, Universidade Federal da Paraíba, João Pessoa, PB, Brazil

Applicability domain

Every prediction is flagged for reliability

The AD histogram shows the distribution of Tanimoto similarities between your molecule and the training set. The colored line marks your molecule's position; the red dashed line marks the 0.40 threshold. Predictions to the left are marked Outside AD.

Active · prob ≥ 70% + IN AD
Borderline · 50–70%
Inactive · prob < 50%
Outside AD · extrapolation