Configuration

Boltz-2

This configuration predicts ligand poses for the SARS-CoV-2 and MERS-CoV main proteases in the ASAP challenge setting.

Sources

A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Conclusions (paragraph 1); Introduction (paragraph 2)

1 evaluation · 1 metric row

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

A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 1); Results/Potency Prediction (paragraph 3)

At a glance

Model type

Biomolecular structure predictor; this record is the paper-specific evaluated configuration.

Sources

jwohlwend/boltz README.md · README.md model description

Biological inputs

Main-protease protein information and ligand structures/SMILES

Sources

A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Methodologies/Pose Prediction/Deep Learning-Based Modeling with DiffDock (paragraph 1); Methodologies/Pose Prediction/Deep Learning-Based Modeling with Boltz-2 (paragraph 1)

Outputs

Predicted ligand binding poses

Sources

A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Methodologies/Pose Prediction/Deep Learning-Based Modeling with Gnina (paragraph 2); Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 2)

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.

Evaluation setupMetricEvidence origin

Results grouped by the exact reported evaluation
Metric and finding	Coverage and uncertainty	Evidence
Boltz-2: Ligand potency prediction using generated poses Configuration: Boltz-2Task: Ligand potency prediction using generated posesDataset: SARS-CoV-2 Mpro ligands Potency prediction using Boltz-2 ligand-pose generation protocol; see paper scoring pipeline. Independent external evaluation · Evaluation metadata: needs review
0.800 Pearson R Unit: unitless · Direction: unknown	Uncertainty: ± 0.027 Scored: Not reported · Eligible: Not reported	source checkedA Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Table 3, Boltz-2 row, Pearson’s R column Source checking is not independent reproduction.

How it works

How the evaluated method works

Boltz-2 is the co-folding pose-generation comparator; downstream LRIP-SF affinity fitting is separate from this row.

Sources

Underlying method and version boundaries

The official Boltz repository publishes separate Boltz-1 and Boltz-2 models. Boltz-2 adds affinity prediction and other changes; these are not retroactively attributed to Boltz-1 evaluations.

Sources

jwohlwend/boltz README.md · README.md; introduction, model description, pretrained-model and usage sections at pinned revision

What was evaluated

The linked evaluation record identifies Boltz-2: Ligand potency prediction using generated poses. Its dataset, split, adaptation and evidence origin remain attached to the reported results.

Sources

A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · The named evaluation’s methods and comparison table; exact preserved evaluation IDs: evaluation-lit-047

Strengths and limitations

Strengths and considerations

Evaluates pose generation and downstream affinity estimation separately.
Sources
A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Conclusions (paragraph 3); Discussion (paragraph 10)

Limitations and conditions

An affinity result obtained by LRIP-SF on a predicted pose must not be assigned to the pose generator alone.
Sources
A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 1); Results/Potency Prediction (paragraph 3)

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-cdc9aabf4efc04

Specifications

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.

Inputs, outputs and configuration
Property	Description and evidence
Model type	Biomolecular structure predictor; this record is the paper-specific evaluated configuration. Sources jwohlwend/boltz README.md · README.md model description
Architecture / procedure	Boltz-2 is the co-folding pose-generation comparator; downstream LRIP-SF affinity fitting is separate from this row. Sources A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 1); Results/Potency Prediction (paragraph 3)
Biological inputs	Main-protease protein information and ligand structures/SMILES Sources A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Methodologies/Pose Prediction/Deep Learning-Based Modeling with DiffDock (paragraph 1); Methodologies/Pose Prediction/Deep Learning-Based Modeling with Boltz-2 (paragraph 1)
Outputs	Predicted ligand binding poses Sources A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Methodologies/Pose Prediction/Deep Learning-Based Modeling with Gnina (paragraph 2); Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 2)
Parameters	An aggregate parameter total for this exact evaluated configuration is not established by the inspected sources. · Not reported in inspected sources Sources (2) A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases; jwohlwend/boltz README.md · Methodologies/Pose Prediction/Data Preparation; Methodologies/Pose Prediction/Molecular Docking with Glide; Methodologies/Pose Prediction/Molecular Docking with AutoDock Vina; Methodologies/Pose Prediction/Flexible Ligand Superposition with FlexS; Methodologies/Pose Prediction/Deep Learning-Based Modeling with AlphaFold3; Methodologies/Pose Prediction/Deep Learning-Based Modeling with DiffDock; Methodologies/Pose Prediction/Deep Learning-Based Modeling with Boltz-2; Methodologies/Pose Prediction/Deep Learning-Based Modeling with Gnina; inspected for aggregate parameter count (component sizes are not added without an exact configuration); README.md at pinned repository revision
Known versions / configuration	Boltz-2 is the comparison-table label; that label does not specify an immutable weight revision. · Not reported in inspected sources Sources A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Model identification in the comparison table and corresponding Methods; immutable checkpoint revision is not supplied by the table label.
Training data / fitting	The study contains 770 SARS-CoV-2 training complexes and test sets of 98 SARS-CoV-2 and 97 MERS-CoV complexes; these are study partitions, not necessarily upstream-model training corpora. Sources A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Methodologies/Pose Prediction/Data Preparation (paragraph 2); Results/Potency Prediction (paragraph 5)
Context limits	A maximum input/context length for this exact evaluated configuration is not established by the inspected sources. · Not reported in inspected sources Sources (2) A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases; jwohlwend/boltz README.md · Methodologies/Pose Prediction/Data Preparation; Methodologies/Pose Prediction/Molecular Docking with Glide; Methodologies/Pose Prediction/Molecular Docking with AutoDock Vina; Methodologies/Pose Prediction/Flexible Ligand Superposition with FlexS; Methodologies/Pose Prediction/Deep Learning-Based Modeling with AlphaFold3; Methodologies/Pose Prediction/Deep Learning-Based Modeling with DiffDock; Methodologies/Pose Prediction/Deep Learning-Based Modeling with Boltz-2; Methodologies/Pose Prediction/Deep Learning-Based Modeling with Gnina; inspected for explicit maximum input length (dataset lengths and family-wide limits are not substituted); README.md at pinned repository revision
Access	Official upstream implementation and usage documentation: https://github.com/jwohlwend/boltz/blob/b1ebfc46ecf57f5414e0d1a6f9027bbb122c53bc/README.md. This pinned documentation revision is not automatically the evaluated weight revision. Sources jwohlwend/boltz README.md · README.md; installation, model download and usage instructions
Code licence	MIT (upstream repository code at the cited revision; this does not establish every dependency or historical checkpoint licence). Sources jwohlwend/boltz LICENSE · LICENSE; complete licence text
Weights licence	MIT for all code and model weights, explicitly stated in the official Boltz README. Historical evaluated weight identity remains separately recorded. Sources jwohlwend/boltz README.md · README.md; license and model release introduction

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.

22 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-09-17-d277315f7d76
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	A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases Original source ↗ Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 1); Results/Potency Prediction (paragraph 3) Version: version of record Retrieved: 2026-09-16T10:41:16.557756+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.caption` Source artifact SHA-256: `c356a1c65a0033e5ae18a05d4afab5495856c5b6869328ff49e13547a4801a57` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram steps ["Main-protease protein information and ligand structures/SMILES","Boltz-2","Predicted ligand binding poses"] Individual claims	A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases Original source ↗ Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 1); Results/Potency Prediction (paragraph 3) Version: version of record Retrieved: 2026-09-16T10:41:16.557756+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.steps` Source artifact SHA-256: `c356a1c65a0033e5ae18a05d4afab5495856c5b6869328ff49e13547a4801a57` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram title Evaluated procedure (conceptual) Individual claims	A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases Original source ↗ Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 1); Results/Potency Prediction (paragraph 3) Version: version of record Retrieved: 2026-09-16T10:41:16.557756+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.diagram.title` Source artifact SHA-256: `c356a1c65a0033e5ae18a05d4afab5495856c5b6869328ff49e13547a4801a57` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Model type Biomolecular structure predictor; this record is the paper-specific evaluated configuration. Individual claims	jwohlwend/boltz README.md Original source ↗ README.md model description Version: b1ebfc46ecf57f5414e0d1a6f9027bbb122c53bc Retrieved: 2026-09-16T20:30:16.575839+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.facts.0.value` Source artifact SHA-256: `79149435313841e6f9cc0c8b581259f3952a2469cbe6f9a733f7143b557025c2` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Architecture / procedure Boltz-2 is the co-folding pose-generation comparator; downstream LRIP-SF affinity fitting is separate from this row. Individual claims	A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases Original source ↗ Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 1); Results/Potency Prediction (paragraph 3) Version: version of record Retrieved: 2026-09-16T10:41:16.557756+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.facts.1.value` Source artifact SHA-256: `c356a1c65a0033e5ae18a05d4afab5495856c5b6869328ff49e13547a4801a57` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Weights licence MIT for all code and model weights, explicitly stated in the official Boltz README. Historical evaluated weight identity remains separately recorded. Individual claims	jwohlwend/boltz README.md Original source ↗ README.md; license and model release introduction Version: b1ebfc46ecf57f5414e0d1a6f9027bbb122c53bc Retrieved: 2026-09-16T20:30:16.575839+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.facts.10.value` Source artifact SHA-256: `79149435313841e6f9cc0c8b581259f3952a2469cbe6f9a733f7143b557025c2` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Biological inputs Main-protease protein information and ligand structures/SMILES Individual claims	A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases Original source ↗ Methodologies/Pose Prediction/Deep Learning-Based Modeling with DiffDock (paragraph 1); Methodologies/Pose Prediction/Deep Learning-Based Modeling with Boltz-2 (paragraph 1) Version: version of record Retrieved: 2026-09-16T10:41:16.557756+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.facts.2.value` Source artifact SHA-256: `c356a1c65a0033e5ae18a05d4afab5495856c5b6869328ff49e13547a4801a57` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Outputs Predicted ligand binding poses Individual claims	A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases Original source ↗ Methodologies/Pose Prediction/Deep Learning-Based Modeling with Gnina (paragraph 2); Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 2) Version: version of record Retrieved: 2026-09-16T10:41:16.557756+00:00	source checked automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.facts.3.value` Source artifact SHA-256: `c356a1c65a0033e5ae18a05d4afab5495856c5b6869328ff49e13547a4801a57` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Parameters An aggregate parameter total for this exact evaluated configuration is not established by the inspected sources. Individual claims	jwohlwend/boltz README.md Original source ↗ Methodologies/Pose Prediction/Data Preparation; Methodologies/Pose Prediction/Molecular Docking with Glide; Methodologies/Pose Prediction/Molecular Docking with AutoDock Vina; Methodologies/Pose Prediction/Flexible Ligand Superposition with FlexS; Methodologies/Pose Prediction/Deep Learning-Based Modeling with AlphaFold3; Methodologies/Pose Prediction/Deep Learning-Based Modeling with DiffDock; Methodologies/Pose Prediction/Deep Learning-Based Modeling with Boltz-2; Methodologies/Pose Prediction/Deep Learning-Based Modeling with Gnina; inspected for aggregate parameter count (component sizes are not added without an exact configuration); README.md at pinned repository revision Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: b1ebfc46ecf57f5414e0d1a6f9027bbb122c53bc Retrieved: 2026-09-16T20:30:16.575839+00:00	unreported automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.facts.4.value` Source artifact SHA-256: `79149435313841e6f9cc0c8b581259f3952a2469cbe6f9a733f7143b557025c2` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Parameters An aggregate parameter total for this exact evaluated configuration is not established by the inspected sources. Individual claims	A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases Original source ↗ Methodologies/Pose Prediction/Data Preparation; Methodologies/Pose Prediction/Molecular Docking with Glide; Methodologies/Pose Prediction/Molecular Docking with AutoDock Vina; Methodologies/Pose Prediction/Flexible Ligand Superposition with FlexS; Methodologies/Pose Prediction/Deep Learning-Based Modeling with AlphaFold3; Methodologies/Pose Prediction/Deep Learning-Based Modeling with DiffDock; Methodologies/Pose Prediction/Deep Learning-Based Modeling with Boltz-2; Methodologies/Pose Prediction/Deep Learning-Based Modeling with Gnina; inspected for aggregate parameter count (component sizes are not added without an exact configuration); README.md at pinned repository revision Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: version of record Retrieved: 2026-09-16T10:41:16.557756+00:00	unreported automated source review · 2026-09-16 Audit 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. Field: `attributes.profile.facts.4.value` Source artifact SHA-256: `c356a1c65a0033e5ae18a05d4afab5495856c5b6869328ff49e13547a4801a57` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact

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

Sources and history

Release 2026-09-17-d277315f7d76 · Record review: needs review

3 source records and release history

A Comparative Study of Deep Learning and Classical Modeling Approaches for Protein–Ligand Binding Pose and Affinity Prediction in Coronavirus Main Proteases · Original source · version of record
jwohlwend/boltz README.md · Original source · b1ebfc46ecf57f5414e0d1a6f9027bbb122c53bc
jwohlwend/boltz LICENSE · Original source · b1ebfc46ecf57f5414e0d1a6f9027bbb122c53bc

Download this release

Technical metadata and extraction receipts

Stable ID: reported-model-cdc9aabf4efc04

areas: molecular-interactions
entity level: method
version: Not reported
reported name: Boltz-2
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: mpro-pose-affinity-2025; evidence-reported-base-boltz-readme-md; source locator: Methodologies/Potency Prediction/LRIP-SF/Pose Generation (paragraph 1); Results/Potency Prediction (paragraph 3) | README.md model description | Conclusions (paragraph 1); Introduction (paragraph 2); 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.

Related records

model: Boltz-2: Ligand potency prediction using generated poses

Suggest a correction