is a freely available web application for simultaneous in silico screening of small molecules against a broad panel of 99 human membrane transporters from the ABC and SLC superfamilies. Comprehensive in silico models capable of predicting the full range of drug–transporter interactions have long remained scarce, while existing computational tools have largely been limited to a narrow set of well-studied proteins, such as P-glycoprotein. The web service was developed in recognition of the critical role that membrane transporters play in pharmacokinetics, pharmacodynamics, and drug–drug interactions.
TIP is built on the PASS algorithm, using Multilevel Neighborhoods of Atoms descriptors and a Bayesian classifier to generate twenty classification SAR knowledge bases from ChEMBL data: one comprehensive base covering all 99 transporters and nineteen specific bases focused on individual, well-studied transporters such as P-glycoprotein, OATP1B1/1B3, and the serotonin, dopamine, and norepinephrine transporters. For transporters with sufficient quantitative data, TIP additionally applies Self-Consistent Regression with Radial Basis Function networks (SCR-RBF) based on Quantitative Neighborhoods of Atoms descriptors, yielding eleven robust quantitative models capable of predicting pIC50 potency values rather than a simple active/inactive classification.
To obtain a prediction, a user enters a chemical structure via a Marvin/Ketcher/JSME applet, SMILES string, or drug name, and selects a model type — Comprehensive SAR, Specific SAR, or Quantitative SAR. The classification models return, for each transporter, the probabilities Pa (probability the compound is an inhibitor) and Pi (probability it is not), along with transporter identifiers (ChEMBL ID, Gene Symbol, UniProt ID, TCDB ID), while the quantitative models output an estimated pIC50 potency value.
Model reliability is evaluated through leave-one-out and 20-fold cross-validation, yielding an average Invariant Accuracy of Prediction (IAP) of 0.95 across all 99 transporters in the Comprehensive model.
A distinctive strength of TIP is its ability to provide both qualitative and quantitative outputs from a single query. This feature is rarely offered by free online tools and substantially improves interpretability for early-stage drug development and transporter-mediated drug-drug interaction profiling.
A distinctive feature of TIP is atom-level color-coded visualization, which highlights the contribution of individual atoms to predicted inhibitory activity using an RGB scheme (green for activity-promoting fragments, red for inactivity-promoting fragments, blue for negligible contribution), helping users rationalize structure–activity relationships directly on the molecule. The service also integrates a UMAP-based chemical space viewer for assessing a compound's position relative to training-set chemistry, scaffold-diversity visualizations for each transporter's known inhibitors, and downloadable SDF training datasets.
TIP is designed to support multiple stages of preclinical drug research and safety assessment:
Early drug candidate screening - rapidly evaluate whether a new compound is likely to inhibit key transporters before committing to costly in vitro or in vivo experiments;
Drug–drug interaction (DDI) profiling - identify potential transporter-mediated DDI liabilities, particularly relevant for drugs co-administered with substrates of P-gp, OATP1B1/1B3, MATE1/2, or OCT2;
Multidrug resistance (MDR) assessment - predict whether a compound may interfere with efflux transporters (e.g., P-gp, BCRP, MRP1) that are overexpressed in tumor cells and associated with chemotherapy resistance;
CNS and neuropharmacology research - assess interactions with monoamine transporters (SERT, DAT, NET), critical for antidepressant, antipsychotic, and stimulant drug design;
Metabolic and renal safety profiling - screen compounds against SGLT1/2 and urate transporters (SLC22A12) to anticipate metabolic or nephrotoxic side effects;
By jointly reporting inhibition probability and predicted potency in a single query, TIP significantly improves the interpretability of results for early-stage drug development and transporter-mediated liability profiling, supporting tasks such as identifying likely drug–drug interaction risks (as demonstrated for ranolazine and febuxostat) and prioritizing compounds with favorable ADME profiles before synthesis or experimental testing. As with other Way2Drug ADME-related web-services, TIP allows researchers to move directly from a structural formula to an actionable pharmacokinetic risk assessment, reducing the time and cost associated with experimental transporter-interaction screening.
Broad coverage in a single query - unlike most freely available tools that focus on one or two transporters, TIP simultaneously screens a compound against up to 99 human transporters, delivering a comprehensive interaction landscape without requiring multiple platforms or paid software.
Quantitative potency estimates alongside qualitative predictions - TIP uniquely combines inhibition probability (Pa/Pi) with predicted pIC₅₀ values, allowing researchers to prioritize not just whether a compound inhibits a transporter, but how potently - a critical distinction for lead optimization.
Regulatory-relevant transporter panel - the service covers the transporter set highlighted by the FDA and International Transporter Consortium for mandatory DDI evaluation (P-gp, BCRP, OATP1B1/1B3, OCT2, MATE1/2), making TIP directly applicable to regulatory submission support and IND-enabling studies.
Anastasia V. Rudik et al. (2026)
TIP: An open-access web resource for transporter interaction prediction.
Computational Toxicology, 39:100438.
If you need to evaluate a large dataset, or if you need to maintain confidentiality of structural formulas transmitted via unsecured data channels, you can contact us to discuss the licensing opportunities.
© 2011 - 2026 Way2Drug. All rights reserved.