Model type
Convolutional neural network; this record is the paper-specific evaluated configuration.
Configuration
AK-score-single
AK-score predicts protein–ligand binding affinity from a three-dimensional complex.
Sources
AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.5. Performance Assessment (paragraph 8); 3. Methods/3.2. Convolutional Neural Network (paragraph 1)How it worksEvaluated procedure (conceptual)
Conceptual input–method–output guide. Check the procedure text and linked evaluation for fitted components, additional inputs and exact settings.
Sources
AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1)At a glance
Biological inputs
Protein–ligand complex coordinates represented as atomic-density grids
Sources
AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.2. Convolutional Neural Network (paragraph 3)Outputs
Predicted binding affinity
Sources
AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.5. Performance Assessment (paragraph 8); 3. Methods/3.5. Performance Assessment (paragraph 10)limited source coverage · Automated source review, 2026-09-16. All specifications and missing details
Evaluations and results
Release 2026-09-17-d277315f7d76 · 1 evaluation · 1 metric row. Different protocols are not a single leaderboard.
| Metric and finding | Coverage and uncertainty | Evidence |
|---|---|---|
| AK-score-single: Protein–ligand binding affinity scoring CASF-2016 scoring-power evaluation. Author-reported evaluation · Evaluation metadata: needs review | ||
| 0.759 Pearson R Unit: unitless · Direction: unknown | Uncertainty: not reported in legacy extract Scored: Not reported · Eligible: Not reported | source checkedAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · Table 2, AK-score-single / learning rate 0.0007 row, Scoring Pearson (R) column Source checking is not independent reproduction. |
How it works
How the evaluated method works
The binding pocket and ligand are voxelised on a 30 Å cube with 1 Å spacing. Multichannel 3D convolutional networks learn interaction patterns; the ensemble averages independently trained models, while the single configuration uses one network.
Sources
AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1)What was evaluated
The linked evaluation record identifies AK-score-single: Protein–ligand binding affinity scoring. Its dataset, split, adaptation and evidence origin remain attached to the reported results.
Sources
AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · The named evaluation’s methods and comparison table; exact preserved evaluation IDs: evaluation-lit-b3-049Strengths and limitations
Strengths and considerations
- The same framework supports affinity regression and relative affinity ranking within the evaluated structural benchmark.
Sources
AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.5. Performance Assessment (paragraph 11); 3. Methods/3.5. Performance Assessment (paragraph 3)
Limitations and conditions
- Requires a complex structure and inherits the chemical and structural coverage of PDBbind; a favourable core-set result does not establish prospective docking accuracy.
Sources
AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · Abstract (paragraph 1); 2. Results and Discussion/2.5. Identifying Hot Spots for Binding Affinity Determination Using Grad-CAM (paragraph 2)
Profile review details
Primary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction.
Stable record: reported-model-de89576d8d316bSpecifications
Inputs, training, access and other details
Explanatory profile: limited source coverage · Automated source review, 2026-09-16. Review applies to the cited claims; unresolved fields are listed below. Numerical results retain their own review status.
| Property | Description and evidence |
|---|---|
| Model type | Convolutional neural network; this record is the paper-specific evaluated configuration.SourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1) |
| Architecture / procedure | The binding pocket and ligand are voxelised on a 30 Å cube with 1 Å spacing. Multichannel 3D convolutional networks learn interaction patterns; the ensemble averages independently trained models, while the single configuration uses one network.SourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1) |
| Biological inputs | Protein–ligand complex coordinates represented as atomic-density gridsSourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.2. Convolutional Neural Network (paragraph 3) |
| Outputs | Predicted binding affinitySourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.5. Performance Assessment (paragraph 8); 3. Methods/3.5. Performance Assessment (paragraph 10) |
| Parameters | Each network has 1,294,925 parameters:1,293,447 trainable and 1,478 non-trainable. The ensemble uses separately trained networks; this is the per-network count.SourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · Methods3.3 Network Architecture; total-parameter paragraph |
| Known versions / configuration | single; learning rate 0.0007SourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · Table ijms-21-08424-t002 (paragraph 1); Table ijms-21-08424-t001 (paragraph 1) |
| Training data / fitting | 3,772 PDBbind-2016 refined-set complexes for training; the 285-complex core set is held out for testing.SourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 2. Results and Discussion/2.4. Assessment with an Additional Dataset (paragraph 1); 3. Methods/3.5. Performance Assessment (paragraph 14) |
| Context limits | 30 × 30 × 30 spatial grid at 1 Å spacingSourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.2. Convolutional Neural Network (paragraph 4) |
| Access | A public release of the original 2020 AK-score model was not established from the paper or targeted official-repository search. AK-Score2 is a separate later method and is not substituted. · Not reported in inspected sourcesSourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · Complete 2020 paper, including Methods and Supplementary Materials statement; exact-name repository discovery |
| Code licence | No explicit code licence was established from the paper’s availability statement and inspected repository-root documentation. · Not reported in inspected sourcesSourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 4. Conclusions (paragraph 1); 3. Methods/3.5. Performance Assessment (paragraph 14) |
| Weights licence | The inspected model-access documentation does not explicitly identify terms for this exact evaluated checkpoint or fitted head; repository code terms are shown separately. · Not reported in inspected sourcesSourcesAK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · 2. Results and Discussion/2.5. Identifying Hot Spots for Binding Affinity Determination Using Grad-CAM (paragraph 1); 4. Conclusions (paragraph 1) |
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.
19 evidence rows matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Diagram caption Conceptual input–method–output guide. Check the procedure text and linked evaluation for fitted components, additional inputs and exact settings. Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1) Version: version of record | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Diagram steps ["Protein–ligand complex coordinates represented as atomic-density grids","AK-score-single","Predicted binding affinity"] Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1) Version: version of record | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Diagram title Evaluated procedure (conceptual) Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1) Version: version of record | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Model type Convolutional neural network; this record is the paper-specific evaluated configuration. Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1) Version: version of record | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Architecture / procedure The binding pocket and ligand are voxelised on a 30 Å cube with 1 Å spacing. Multichannel 3D convolutional networks learn interaction patterns; the ensemble averages independently trained models, while the single configuration uses one network. Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1) Version: version of record | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Weights licence The inspected model-access documentation does not explicitly identify terms for this exact evaluated checkpoint or fitted head; repository code terms are shown separately. Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks 2. Results and Discussion/2.5. Identifying Hot Spots for Binding Affinity Determination Using Grad-CAM (paragraph 1); 4. Conclusions (paragraph 1) Version: version of record | unreported automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Biological inputs Protein–ligand complex coordinates represented as atomic-density grids Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.2. Convolutional Neural Network (paragraph 3) Version: version of record | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Outputs Predicted binding affinity Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks 3. Methods/3.5. Performance Assessment (paragraph 8); 3. Methods/3.5. Performance Assessment (paragraph 10) Version: version of record | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Parameters Each network has 1,294,925 parameters:1,293,447 trainable and 1,478 non-trainable. The ensemble uses separately trained networks; this is the per-network count. Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks Methods3.3 Network Architecture; total-parameter paragraph Version: version of record | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
| Known versions / configuration single; learning rate 0.0007 Individual claims | AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks Table ijms-21-08424-t002 (paragraph 1); Table ijms-21-08424-t001 (paragraph 1) Version: version of record | source checked automated source review · 2026-09-16 Audit detailsPrimary full text and the available official implementation/model documentation were inspected. Explanatory claims are source-backed; unresolved exact-configuration metadata is labelled explicitly. This is automated review, not a human review or independent benchmark reproduction. Field: Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
Sources and history
Release 2026-09-17-d277315f7d76 · Record review: needs review
1 source records and release history
- AK-Score: Accurate Protein-Ligand Binding Affinity Prediction Using an Ensemble of 3D-Convolutional Neural Networks · Original source · version of record
Technical metadata and extraction receipts
Stable ID: reported-model-de89576d8d316b
- areas
- molecular-interactions
- entity level
- method
- version
- single; learning rate 0.0007
- reported name
- AK-score-single
- historical missing metadata
- checkpoint revision: not_reported_in_legacy_extract; training data: not_reported_in_legacy_extract; licence: not_reported_in_legacy_extract
- metadata review scope
- historical_missing_metadata preserves the original discovery state. Current descriptive evidence and missingness are recorded in profile.facts; numerical-result review is separate.
- legacy kinds
- model
- entity classification
- review date: 2026-09-17; rationale: This source-scoped entry preserves the method/configuration actually named in an evaluation. It is neither a global family identity nor proof of an immutable checkpoint; the linked evaluation retains adaptation, fitting and scoring details.; source ids: akscore-2020; source locator: 3. Methods/3.2. Convolutional Neural Network (paragraph 1); 3. Methods/3.4. Ensemble Prediction (paragraph 1) | 3. Methods/3.5. Performance Assessment (paragraph 8); 3. Methods/3.2. Convolutional Neural Network (paragraph 1); ambiguities: Configuration means the source-labelled evaluated identity. It does not establish missing checkpoint hashes, default settings or equivalence to same-named records in other papers.