Model type
Study-specific predictive method; this record is the paper-specific evaluated configuration.
Configuration
TransBind
TransBind predicts transcription-factor binding by combining genomic DNA with protein-aware TF embeddings.
Sources
Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Ablation study/Multimodal fusion strategies (paragraph 1); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5)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
Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4)At a glance
Biological inputs
DNA windows and transcription-factor protein representations
Sources
Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Ablation study/Effect of input DNA sequence length on model performance (paragraph 3); Materials and methods/Deep learning model for label-zero-shot TF–DNA binding site prediction (paragraph 3)Outputs
TF-binding predictions for genomic regions
Sources
Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Introduction (paragraph 8); Materials and methods/DNA data (paragraph 4)limited source coverage · Automated source review, 2026-09-16. All specifications and missing details
Evaluations and results
Release 2026-09-17-d277315f7d76 · 1 evaluation · 2 metric rows. Different protocols are not a single leaderboard.
| Metric and finding | Coverage and uncertainty | Evidence |
|---|---|---|
| TransBind: transcription-factor DNA binding-site prediction Configuration: TransBindProtocol: TF-cell-type binding on held-out chromosomes8and9 (transcription-factor DNA binding-site prediction)Dataset: genome-wide TF binding sites Per-label AUROC/AUPR macro-averaged over 690 labels. Same Table 2 test cohort; methods differ in additional protein/dynamics inputs. DeepSEA-derived 690 TF-cell-type labels; chr 7 validation; remaining autosomes+X training; chr 8/9 test. Author-reported evaluation · Evaluation metadata: needs review | ||
| 0.9508 AUROC Unit: fraction · Direction: higher | Uncertainty: not reported in legacy extract Scored: Not reported · Eligible: Not reported | source checkedIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction; Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Table 2, TransBind row, AUROC column Source checking is not independent reproduction. |
| 0.3741 AUPR Unit: fraction · Direction: higher | Uncertainty: unreported Scored: Not reported · Eligible: Not reported | source checkedIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Table 2., row TransBind, column AUPR; XML row7 column3 Source checking is not independent reproduction. |
How it works
How the evaluated method works
Cross-attention allows a TF embedding containing sequence and structural information to attend to DNA regions. The protein language model is pretrained on DNA-binding proteins.
Sources
Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4)What was evaluated
The linked evaluation record identifies TransBind: transcription-factor DNA binding-site prediction. Its dataset, split, adaptation and evidence origin remain attached to the reported results.
Sources
Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · The named evaluation’s methods and comparison table; exact preserved evaluation IDs: evaluation-lit-b4-023Strengths and limitations
Strengths and considerations
- Makes TF identity and representation explicit instead of using DNA alone.
Sources
Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Results/Convolutional kernels capture general DNA patterns (paragraph 4); Ablation study/Effect of input DNA sequence length on model performance (paragraph 3)
Limitations and conditions
- ChIP-seq labels, cell context and unseen-TF partitioning define the biological claim; protein-aware prediction does not itself confirm binding in a new experiment.
Sources
Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Introduction (paragraph 8); Abstract (paragraph 1)
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-81b0394d5ac3e8Specifications
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 | Study-specific predictive method; this record is the paper-specific evaluated configuration.SourcesIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4) |
| Architecture / procedure | Cross-attention allows a TF embedding containing sequence and structural information to attend to DNA regions. The protein language model is pretrained on DNA-binding proteins.SourcesIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4) |
| Biological inputs | DNA windows and transcription-factor protein representationsSourcesIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Ablation study/Effect of input DNA sequence length on model performance (paragraph 3); Materials and methods/Deep learning model for label-zero-shot TF–DNA binding site prediction (paragraph 3) |
| Outputs | TF-binding predictions for genomic regionsSourcesIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Introduction (paragraph 8); Materials and methods/DNA data (paragraph 4) |
| Parameters | 4.6 million parameters for the selected unidirectional cross-attention configuration; this does not count the upstream pretrained feature extractor.SourcesIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Ablation study/Multimodal fusion strategies (paragraph 1); Ablation study/Multimodal fusion strategies (paragraph 2) |
| Known versions / configuration | TransBind is the comparison-table label; that label does not specify an immutable weight revision. · Not reported in inspected sourcesSourcesIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Model identification in the comparison table and corresponding Methods; immutable checkpoint revision is not supplied by the table label. |
| Training data / fitting | 690 ChIP-seq experiments covering 161 TFs and 91 human cell types; the reported final DNA-bin collection contains 1,903,668 unique unambiguous bins before reverse-complement augmentation.SourcesIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Materials and methods/DNA data (paragraph 3); Materials and methods/DNA data (paragraph 1) |
| Context limits | 1,000-bp inputs: a 200-bp genomic bin extended by 400 bases on each side. The assembly is GRCh37/hg19; chromosomes 8–9 are test and chromosome 7 is validation.SourcesIntegrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Materials and methods/DNA data (paragraph 2); Table tbl1 (paragraph 1) |
| Access | Official study implementation and usage documentation: https://github.com/jianlin-cheng/TransBind/blob/7537f264c5ad94958bcad05bb57edd8028c323ff/README.md. This pinned documentation revision is not automatically the evaluated weight revision.Sourcesjianlin-cheng/TransBind README.md · README.md; installation, model download and usage instructions |
| Code licence | GNU GPL version 3 (study repository code at the cited revision; this does not establish every dependency or historical checkpoint licence).Sourcesjianlin-cheng/TransBind LICENSE · LICENSE; complete licence text |
| 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 sourcesSourcesjianlin-cheng/TransBind README.md · README.md; checkpoint/access documentation and licence scope |
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 | Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4) 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 ["DNA windows and transcription-factor protein representations","TransBind","TF-binding predictions for genomic regions"] Individual claims | Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4) 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 | Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4) 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 Study-specific predictive method; this record is the paper-specific evaluated configuration. Individual claims | Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4) 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 Cross-attention allows a TF embedding containing sequence and structural information to attend to DNA regions. The protein language model is pretrained on DNA-binding proteins. Individual claims | Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4) 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 | jianlin-cheng/TransBind README.md README.md; checkpoint/access documentation and licence scope Version: 7537f264c5ad94958bcad05bb57edd8028c323ff | 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 DNA windows and transcription-factor protein representations Individual claims | Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction Ablation study/Effect of input DNA sequence length on model performance (paragraph 3); Materials and methods/Deep learning model for label-zero-shot TF–DNA binding site prediction (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 TF-binding predictions for genomic regions Individual claims | Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction Introduction (paragraph 8); Materials and methods/DNA data (paragraph 4) 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 4.6 million parameters for the selected unidirectional cross-attention configuration; this does not count the upstream pretrained feature extractor. Individual claims | Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction Ablation study/Multimodal fusion strategies (paragraph 1); Ablation study/Multimodal fusion strategies (paragraph 2) 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 TransBind is the comparison-table label; that label does not specify an immutable weight revision. Individual claims | Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction Model identification in the comparison table and corresponding Methods; immutable checkpoint revision is not supplied by the table label. 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 |
Sources and history
Release 2026-09-17-d277315f7d76 · Record review: needs review
3 source records and release history
- Integrating protein and DNA embeddings for improving genome-wide transcription factor binding site prediction · Original source · version of record
- jianlin-cheng/TransBind README.md · Original source · 7537f264c5ad94958bcad05bb57edd8028c323ff
- jianlin-cheng/TransBind LICENSE · Original source · 7537f264c5ad94958bcad05bb57edd8028c323ff
Technical metadata and extraction receipts
Stable ID: reported-model-81b0394d5ac3e8
- areas
- molecular-interactions
- entity level
- method
- version
- Not reported
- reported name
- TransBind
- historical missing metadata
- version: not_reported_in_legacy_extract; 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: transbind-2026; source locator: Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 3: Bimodal feature aggregation for TF–DNA binding prediction (paragraph 4) | Ablation study/Multimodal fusion strategies (paragraph 1); Materials and methods/Deep learning model for TF–DNA binding site classification/Module 1: DNA sequence encoder (paragraph 5); 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.