Model type
Two 1D convolutional encoders with a fully connected regression head
Model
DeepDTA
DeepDTA predicts drug-target binding affinity from a ligand SMILES string and a protein sequence, using a separate 1D convolutional encoder for each.
Sources
DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 abstract, p.1; Proposed model and Figure 2, PDF pp.7-8Overview
Inputs
Ligand SMILES string and protein amino-acid sequence
Sources
DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 abstract, p.1; Proposed model, PDF p.7Output
Continuous binding affinity, trained with mean squared error
Sources
DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 Proposed model, PDF pp.7-8limited source coverage · Automated source review, 2026-09-24. All specifications and missing details
Evaluations and results
4 evaluations · 4 metric rows. Different protocols are not a single leaderboard.
Filter evaluations
Applied filters: All linked evaluations
| Tested configuration | Protocol and dataset | Finding | Evidence and details |
|---|---|---|---|
| Configuration: DeepDTA (ATOM3D baseline) (cited as [Öztürk et al., 2018]) | Task: ATOM3D LBA-RMSE: Ligand binding affinity, root mean squared error Dataset subset: ATOM3D LBA (ATOM3D split) | 1.56 rmse error · lower Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · source checkedMethods, coverage and sourceDeepDTA (ATOM3D baseline) on ATOM3D LBA-RMSE: Ligand binding affinity, root mean squared error Trained and scored under the ATOM3D split for this task. Asterisks in the paper mark a run whose training data differed. Aggregation: Not reported ATOM3D: Tasks On Molecules in Three Dimensions · Table 5, row(LBA RMSE), column([Öztürk et al., 2018]) |
| Configuration: DeepDTA (ATOM3D baseline) (cited as [Öztürk et al., 2018]) | Task: ATOM3D LBA-RP: Ligand binding affinity, global Pearson correlation Dataset subset: ATOM3D LBA (ATOM3D split) | 0.573 pearson_r correlation · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · source checkedMethods, coverage and sourceDeepDTA (ATOM3D baseline) on ATOM3D LBA-RP: Ligand binding affinity, global Pearson correlation Trained and scored under the ATOM3D split for this task. Asterisks in the paper mark a run whose training data differed. Aggregation: Not reported ATOM3D: Tasks On Molecules in Three Dimensions · Table 5, row(glob. RP), column([Öztürk et al., 2018]) |
| Configuration: DeepDTA (ATOM3D baseline) (cited as [Öztürk et al., 2018]) | Task: ATOM3D LBA-RS: Ligand binding affinity, global Spearman correlation Dataset subset: ATOM3D LBA (ATOM3D split) | 0.574 spearman_r correlation · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · source checkedMethods, coverage and sourceDeepDTA (ATOM3D baseline) on ATOM3D LBA-RS: Ligand binding affinity, global Spearman correlation Trained and scored under the ATOM3D split for this task. Asterisks in the paper mark a run whose training data differed. Aggregation: Not reported ATOM3D: Tasks On Molecules in Three Dimensions · Table 5, row(glob. RS), column([Öztürk et al., 2018]) |
| Configuration: DeepDTA (ATOM3D baseline) (cited as [Öztürk et al., 2018]) | Task: ATOM3D LEP-AUROC: Ligand efficacy prediction Dataset subset: ATOM3D LEP (ATOM3D split) | 0.696 auroc fraction · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · source checkedMethods, coverage and sourceDeepDTA (ATOM3D baseline) on ATOM3D LEP-AUROC: Ligand efficacy prediction Trained and scored under the ATOM3D split for this task. Asterisks in the paper mark a run whose training data differed. Aggregation: Not reported ATOM3D: Tasks On Molecules in Three Dimensions · Table 5, row(LEP AUROC), column([Öztürk et al., 2018]) |
Source checking is not independent reproduction. Release 2026-09-24-eb3fb1cb4c7f.
Use this model
How it works, versions and access
Versions and evaluated configurations
- DeepDTA (ATOM3D baseline) · family
How it works
How it works
SMILES strings and protein sequences are label-encoded characters. Each passes through its own block of three 1D convolutional layers, where the second and third layers have two and three times the filters of the first, followed by max-pooling. The two pooled vectors are concatenated and fed into fully connected layers of 1,024, 1,024 and 512 units, with dropout after the first two, and a regression output trained with mean squared error.
Sources
DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 input representation, PDF p.4; Proposed model and Figure 2, PDF pp.7-8Original evaluation
The paper evaluates on the Davis kinase dataset (Kd) and the KIBA dataset. It fixes maximum lengths of 85 SMILES and 1,200 protein characters for Davis and 100 and 1,000 for KIBA.
Sources
DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 Datasets and Table 1, PDF p.3; input representation, PDF p.6Strengths, limitations and unresolved questions
Strengths and limitations
Strengths and considerations
No source-reviewed explanatory claims are recorded here yet.
Limitations and conditions
- Uses only the 1D sequence and SMILES strings, not the 3D structure of the complex.
Sources
DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 abstract, p.1
Profile review details
AI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction.
Stable record: identity-model-deepdtaSpecifications
Inputs, training, access and other details
Explanatory profile: limited source coverage · Automated source review, 2026-09-24. Review applies to the cited claims; unresolved fields are listed below. Numerical results retain their own review status.
| Property | Description and evidence |
|---|---|
| Model type | Two 1D convolutional encoders with a fully connected regression headSourcesDeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 Proposed model and Figure 2, PDF pp.7-8 |
| Inputs | Ligand SMILES string and protein amino-acid sequenceSourcesDeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 abstract, p.1; Proposed model, PDF p.7 |
| Output | Continuous binding affinity, trained with mean squared errorSourcesDeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 Proposed model, PDF pp.7-8 |
| Original datasets | Davis (Kd) and KIBASourcesDeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 Datasets and Table 1, PDF p.3 |
| Code | github.com/hkmztrk/DeepDTASourcesDeepDTA repository README (hkmztrk/DeepDTA) · README at a546a8433a6822e958f36171c4356ad6f414d623 |
| Parameter count | Not extracted · Needs further source reviewSourcesDeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · arXiv:1801.10193v2 |
| Known versions | Not extracted or verified for this record. |
| Training data | Not extracted or verified for this record. |
| Context limits | Not extracted or verified for this record. |
| Access | Not extracted or verified for this record. |
| Code licence | Not extracted or verified for this record. |
| Weights licence | Not extracted or verified for this record. |
Evidence
Source checking verifies the cited claim or transcription. It does not establish independent reproduction.
Evidence table
Inspect claims, sources and review details
Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.
One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.
10 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Model type Two 1D convolutional encoders with a fully connected regression head Individual claims | DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) arXiv:1801.10193v2 Proposed model and Figure 2, PDF pp.7-8 Version: 1801.10193v2 | source checked automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
| Inputs Ligand SMILES string and protein amino-acid sequence Individual claims | DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) arXiv:1801.10193v2 abstract, p.1; Proposed model, PDF p.7 Version: 1801.10193v2 | source checked automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
| Output Continuous binding affinity, trained with mean squared error Individual claims | DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) arXiv:1801.10193v2 Proposed model, PDF pp.7-8 Version: 1801.10193v2 | source checked automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
| Original datasets Davis (Kd) and KIBA Individual claims | DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) arXiv:1801.10193v2 Datasets and Table 1, PDF p.3 Version: 1801.10193v2 | source checked automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
| Code github.com/hkmztrk/DeepDTA Individual claims | DeepDTA repository README (hkmztrk/DeepDTA) README at a546a8433a6822e958f36171c4356ad6f414d623 Version: a546a8433a6822e958f36171c4356ad6f414d623 | source checked automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
| Parameter count Not extracted Individual claims | DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) arXiv:1801.10193v2 Version: 1801.10193v2 | unextracted automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
| Limitation Uses only the 1D sequence and SMILES strings, not the 3D structure of the complex. Individual claims | DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) arXiv:1801.10193v2 abstract, p.1 Version: 1801.10193v2 | source checked automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
| How it works SMILES strings and protein sequences are label-encoded characters. Each passes through its own block of three 1D convolutional layers, where the second and third layers have two and three times the filters of the first, followed by max-pooling. The two pooled vectors are concatenated and fed into fully connected layers of 1,024, 1,024 and 512 units, with dropout after the first two, and a regression output trained with mean squared error. Individual claims | DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) arXiv:1801.10193v2 input representation, PDF p.4; Proposed model and Figure 2, PDF pp.7-8 Version: 1801.10193v2 | source checked automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
| Original evaluation The paper evaluates on the Davis kinase dataset (Kd) and the KIBA dataset. It fixes maximum lengths of 85 SMILES and 1,200 protein characters for Davis and 100 and 1,000 for KIBA. Individual claims | DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) arXiv:1801.10193v2 Datasets and Table 1, PDF p.3; input representation, PDF p.6 Version: 1801.10193v2 | source checked automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
| Introduction DeepDTA predicts drug-target binding affinity from a ligand SMILES string and a protein sequence, using a separate 1D convolutional encoder for each. Individual claims | DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) arXiv:1801.10193v2 abstract, p.1; Proposed model and Figure 2, PDF pp.7-8 Version: 1801.10193v2 | source checked automated source review · 2026-09-24 Audit detailsAI-assisted review of the cited claims against pinned primary sources. No human scientific review and no independent reproduction. Field: Source artifact SHA-256: Hash scope: SHA-256 of the retrieved original artifact bytes |
Sources and history
View linked audit checks and correction history
Release 2026-09-24-eb3fb1cb4c7f · Record review: source checked
2 source records and release history
- DeepDTA: Deep Drug-Target Binding Affinity Prediction (arXiv:1801.10193v2) · Original source · 1801.10193v2
- DeepDTA repository README (hkmztrk/DeepDTA) · Original source · a546a8433a6822e958f36171c4356ad6f414d623
Technical metadata and extraction receipts
Stable ID: identity-model-deepdta
- areas
- molecular-interactions
- entity level
- method
- reported name
- DeepDTA
- version
- Not reported
- entity classification
- review date: 2026-09-24; rationale: The cited paper names DeepDTA as a learned drug-target binding affinity predictor. This record covers the method; no checkpoint is implied.; source ids: source-label-deepdta-arxiv-3c368fed; source locator: arXiv:1801.10193v2 title and abstract, p.1; Proposed model and Figure 2, PDF pp.7-8; ambiguities: The paper trains separate models for Davis and KIBA with different input lengths. No common checkpoint is inferred for configurations that retrain DeepDTA.
Related records
- family: DeepDTA (ATOM3D baseline)