Configuration

DeepAffinity (unified RNN/RNN-CNN; DSSP-derived SPS)

DeepAffinity predicts compound-protein affinity from a compound SMILES string and a structure property-annotated protein sequence (SPS), using pretrained recurrent seq2seq encoders followed by convolutional layers.

Identified as DeepAffinity. The source table prints [Karimi et al., 2019].

Configuration: Unified RNN/RNN-CNN retrained by the ATOM3D authors for LBA, with structure annotations computed by DSSP from 3D structures.

ATOM3D §5.1 compares to 'DeepAffinity [Karimi et al., 2019]' for LBA, and App. F.5 specifies 'their unified RNN/RNN-CNN model' with SMILES and SPS inputs and links the public repository. The cited paper introduces this unified RNN-CNN as DeepAffinity.

SPS annotations: Secondary structure and relative solvent accessibility computed with DSSP from the 3D structure, not predicted with SSpro/ACCpro as in the original paper.
Input lengths: Maximum SMILES and SPS lengths raised from 100 and 152 to 160 and 168.
Training: Pretrained seq2seq encoders initialise joint supervised training of the encoders and CNN for 1000 epochs. Other hyperparameters follow the original paper.
Tasks: LBA only. Table 5 has no DeepAffinity LEP value.

Not established: LBA split. Table 8 has no DeepAffinity row, so the identity split behind the Table 5 values is not stated. Attention and ensemble choice. The original paper reports separate, marginalized and joint attention variants. ATOM3D does not say which was used. Checkpoint. No retrained ATOM3D weights or checkpoint are identified.

ATOM3D: Tasks On Molecules in Three Dimensions; DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2); DeepAffinity repository README (Shen-Lab/DeepAffinity) · ATOM3D arXiv:2012.04035v4 §5.1 and Table 5, p.7; App. F.5 with footnote 11, p.24. DeepAffinity arXiv:1806.07537v2 title, p.1; §2.2.2, p.3; §§2.4-2.5, p.4. github.com/Shen-Lab/DeepAffinity README at debca4c9f01991a37594e1b67f3dbb46a29e82d1.. AI-assisted source review on 2026-09-24; no human scientific review.

Sources

DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 abstract, p.1; §2.2.2, p.3; §2.4, p.4

3 evaluations · 3 results

Overview

Model type

Unified RNN-CNN: GRU seq2seq encoders with a convolutional regression head

Sources

DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §§2.3-2.4, p.4

Inputs

Compound SMILES string and protein SPS string

Sources

DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §2.2, PDF pp.2-3

Output

Affinity on a logarithmic scale (pIC50, pKi or pKd)

Sources

DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §2.1, p.2

limited source coverage · Automated source review, 2026-09-24. All specifications and missing details

Evaluations and results

3 evaluations · 3 results. Different protocols are not a single leaderboard.

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: DeepAffinity (unified RNN/RNN-CNN; DSSP-derived SPS) (cited as [Karimi et al., 2019])	Task: ATOM3D LBA-RMSE: Ligand binding affinity, root mean squared error Dataset subset: ATOM3D LBA (ATOM3D split)	1.89 rmse error · lower Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source DeepAffinity (unified RNN/RNN-CNN; DSSP-derived SPS) 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([Karimi et al., 2019])
Configuration: DeepAffinity (unified RNN/RNN-CNN; DSSP-derived SPS) (cited as [Karimi et al., 2019])	Task: ATOM3D LBA-RP: Ligand binding affinity, global Pearson correlation Dataset subset: ATOM3D LBA (ATOM3D split)	0.415 pearson_r correlation · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source DeepAffinity (unified RNN/RNN-CNN; DSSP-derived SPS) 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([Karimi et al., 2019])
Configuration: DeepAffinity (unified RNN/RNN-CNN; DSSP-derived SPS) (cited as [Karimi et al., 2019])	Task: ATOM3D LBA-RS: Ligand binding affinity, global Spearman correlation Dataset subset: ATOM3D LBA (ATOM3D split)	0.426 spearman_r correlation · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source DeepAffinity (unified RNN/RNN-CNN; DSSP-derived SPS) 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([Karimi et al., 2019])

Source checking is not independent reproduction. Release 2026-09-29-06401fd5b220.

Use this model

How it works, versions and access

Related profile: DeepAffinity. This page retains the exact record and its evaluation context.

This configuration

DeepAffinity, over ligand SMILES and annotated protein sequence, as the ATOM3D text describes it.

record: DeepAffinity (unified RNN/RNN-CNN; DSSP-derived SPS)
configuration: Not reported
entity type: Configuration

How it works

GRU seq2seq auto-encoders are first trained without labels on compound SMILES and protein SPS strings. In the unified RNN-CNN model, a 1D convolution and max-pooling layer is added after each encoder, the two outputs are concatenated and passed through two fully connected layers, and the whole pipeline is trained end to end with the pretrained encoders as initialisation. A separate RNN-CNN baseline keeps the encoders fixed.

Sources

DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §§2.3-2.4, p.4

Protein representation

In the original paper, SPS strings group residues into secondary structure elements using secondary structure and solvent accessibility predicted from sequence by SSpro/ACCpro.

Sources

DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §2.2.2, p.3

Variants

The paper also describes separate, marginalized and joint attention mechanisms trained jointly with the encoders and CNN, and a unified RNN/GCNN-CNN variant that replaces the compound RNN with a graph CNN.

Sources (2)

DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2); DeepAffinity repository README (Shen-Lab/DeepAffinity) · arXiv:1806.07537v2 §2.5, p.4 and PDF p.8; README Table of contents

Strengths, limitations and unresolved questions

Strengths and limitations

Strengths and considerations

No source-reviewed explanatory claims are recorded here yet.

Limitations and conditions

In the original paper the protein structure annotations are predicted from sequence rather than taken from solved structures.
Sources
DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §2.2.2, p.3

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-deepaffinity

Specifications

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.

Inputs, outputs and configuration
Property	Description and evidence
Model type	Unified RNN-CNN: GRU seq2seq encoders with a convolutional regression head Sources DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §§2.3-2.4, p.4
Inputs	Compound SMILES string and protein SPS string Sources DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §2.2, PDF pp.2-3
Output	Affinity on a logarithmic scale (pIC50, pKi or pKd) Sources DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §2.1, p.2
Original training data	BindingDB IC50, Ki and Kd labels, with four protein classes held out of IC50 training Sources DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2 §2.1, p.2
Code	github.com/Shen-Lab/DeepAffinity Sources DeepAffinity repository README (Shen-Lab/DeepAffinity) · README at debca4c9f01991a37594e1b67f3dbb46a29e82d1
Parameter count	Not extracted · Needs further source review Sources DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · arXiv:1806.07537v2
Known versions	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.

2 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-09-29-06401fd5b220
Property and statement	Original source and location	Review and provenance
Relationship: family identity-model-deepaffinity Individual claims	ATOM3D: Tasks On Molecules in Three Dimensions Original source ↗ ATOM3D §5.1, p.7; App. F.5, p.24. DeepAffinity arXiv:1806.07537v2 §2.4, p.4. Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Primary full-text snapshot retrieved 2026-09-17; exact bytes pinned by SHA-256 Retrieved: 2026-09-17T08:06:28.182997+00:00	source checked automated source review · 2026-09-24 Audit details Source review establishes this relationship only. Exact evaluated configurations and original numerical review status remain unchanged. ATOM3D retrained the unified RNN/RNN-CNN DeepAffinity model with DSSP-derived SPS input. Family attribution only: the retrained weights and their LBA scores stay with this configuration. Field: `links:family:identity-model-deepaffinity` Claim: model-evaluation-identity-f7f525c3a659cd81f240 Source artifact SHA-256: `92656c20a15311c32bed9edc7f465bb26eb30338f40324fe43bec4b1fc6a7890` Hash scope: Exact retrieved primary paper artifact bytes. Inspected artifact
Relationship: family identity-model-deepaffinity Individual claims	DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) Original source ↗ ATOM3D §5.1, p.7; App. F.5, p.24. DeepAffinity arXiv:1806.07537v2 §2.4, p.4. Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: 1806.07537v2 Retrieved: 2026-09-24T19:28:11+00:00	source checked automated source review · 2026-09-24 Audit details Source review establishes this relationship only. Exact evaluated configurations and original numerical review status remain unchanged. ATOM3D retrained the unified RNN/RNN-CNN DeepAffinity model with DSSP-derived SPS input. Family attribution only: the retrained weights and their LBA scores stay with this configuration. Field: `links:family:identity-model-deepaffinity` Claim: model-evaluation-identity-f7f525c3a659cd81f240 Source artifact SHA-256: `97d6e3ae6a1ac96a5b98ebca778ab0608ab921e71de6e948e97628c44053534d` Hash scope: SHA-256 of the retrieved original artifact bytes Inspected artifact

All evidence (CSV)All evidence (JSONL)How to read the evidence

Sources and history

View linked audit checks and correction history

Release 2026-09-29-06401fd5b220 · Record review: source checked

3 source records and release history

ATOM3D: Tasks On Molecules in Three Dimensions · Original source · Primary full-text snapshot retrieved 2026-09-17; exact bytes pinned by SHA-256
DeepAffinity: interpretable deep learning of compound–protein affinity (arXiv:1806.07537v2) · Original source · 1806.07537v2
DeepAffinity repository README (Shen-Lab/DeepAffinity) · Original source · debca4c9f01991a37594e1b67f3dbb46a29e82d1

Download this release

Technical metadata and extraction receipts

Stable ID: atom3d-method-karimi-et-al-2019

areas: molecular-interactions
source locator: Table 5, column([Karimi et al., 2019])
missing metadata: checkpoint revision: unreported; parameters: unextracted
source label: [Karimi et al., 2019]
source identity: status: resolved; label form: bracketed_citation; display name: DeepAffinity (unified RNN/RNN-CNN; DSSP-derived SPS); identity: DeepAffinity; configuration: Unified RNN/RNN-CNN retrained by the ATOM3D authors for LBA, with structure annotations computed by DSSP from 3D structures.; basis: ATOM3D §5.1 compares to 'DeepAffinity [Karimi et al., 2019]' for LBA, and App. F.5 specifies 'their unified RNN/RNN-CNN model' with SMILES and SPS inputs and links the public repository. The cited paper introduces this unified RNN-CNN as DeepAffinity.; source ids: evidence-expansion-atom3d-92656c20; source-label-deepaffinity-arxiv-97d6e3ae; source-label-deepaffinity-readme-debca4c9; source locator: ATOM3D arXiv:2012.04035v4 §5.1 and Table 5, p.7; App. F.5 with footnote 11, p.24. DeepAffinity arXiv:1806.07537v2 title, p.1; §2.2.2, p.3; §§2.4-2.5, p.4. github.com/Shen-Lab/DeepAffinity README at debca4c9f01991a37594e1b67f3dbb46a29e82d1.; known details: label: SPS annotations; value: Secondary structure and relative solvent accessibility computed with DSSP from the 3D structure, not predicted with SSpro/ACCpro as in the original paper.; source ids: evidence-expansion-atom3d-92656c20; source-label-deepaffinity-arxiv-97d6e3ae; source locator: ATOM3D App. F.5, p.24; DeepAffinity §2.2.2, p.3; label: Input lengths; value: Maximum SMILES and SPS lengths raised from 100 and 152 to 160 and 168.; source ids: evidence-expansion-atom3d-92656c20; source locator: ATOM3D App. F.5, p.24; label: Training; value: Pretrained seq2seq encoders initialise joint supervised training of the encoders and CNN for 1000 epochs. Other hyperparameters follow the original paper.; source ids: evidence-expansion-atom3d-92656c20; source locator: ATOM3D App. F.5, p.24; label: Tasks; value: LBA only. Table 5 has no DeepAffinity LEP value.; source ids: evidence-expansion-atom3d-92656c20; source locator: ATOM3D Table 5, p.7; unknown: label: LBA split; note: Table 8 has no DeepAffinity row, so the identity split behind the Table 5 values is not stated.; label: Attention and ensemble choice; note: The original paper reports separate, marginalized and joint attention variants. ATOM3D does not say which was used.; label: Checkpoint; note: No retrained ATOM3D weights or checkpoint are identified.; original name: [Karimi et al., 2019]; original description: DeepAffinity, over ligand SMILES and annotated protein sequence, as the ATOM3D text describes it.; review: method: automated_source_review; date: 2026-09-24; note: AI-assisted review against the cited primary sources. No human scientific review. Values, locators and comparison conditions are unchanged.

Related records

Suggest a correction