Configuration

ProkBERT-mini

ProkBERT-mini learns microbial DNA representations with local-context-aware tokenisation.

Sources

ProkBERT family: genomic language models for microbiome applications · 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 4 Conclusion (paragraph 5)

1 evaluation · 4 results

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

Overview

Model type

Transformer representation pipeline; this record is the paper-specific evaluated configuration.

Sources

Biological inputs

Microbial DNA sequences

Sources

ProkBERT family: genomic language models for microbiome applications · 1 Introduction (paragraph 7); 4 Conclusion (paragraph 10)

Outputs

DNA embeddings and task-specific genomic classifications

Sources

ProkBERT family: genomic language models for microbiome applications · 2 Materials and methods/2.5 Applied metrics (paragraph 1); 3 Results and discussion/3.1 ProkBERT's learned representations capture genomic structure and phylogeny (paragraph 4)

Access

Official study implementation and usage documentation: https://github.com/nbrg-ppcu/prokbert/blob/8670ae92b816cff158a0b85647a8dea122e251eb/README.md. This pinned documentation revision is not automatically the evaluated weight revision.

Sources

nbrg-ppcu/prokbert README.md · README.md; installation, model download and usage instructions

limited source coverage · Automated source review, 2026-09-16. All specifications and missing details

Evaluations and results

1 evaluation · 4 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: ProkBERT-mini	Protocol: E. coli sigma70 independent promoter test (E. coli sigma70 promoter prediction) Dataset: E. coli sigma70 promoter dataset	0.87 Accuracy unitless · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source ProkBERT-mini: E. coli sigma70 promoter prediction Test-only evaluation on Cassiano and Silva-Rocha 2020 data; methods have different training histories. 865 high-evidence RegulonDB 10.5 promoters and 1,000 nucleotide-distribution-matched negative sequences. Promoter exact matches removed from model training. Aggregation: Not reported ProkBERT family: genomic language models for microbiome applications; ProkBERT family: genomic language models for microbiome applications · Table 3, ProkBERT-mini row, Accuracy column
Configuration: ProkBERT-mini	Protocol: E. coli sigma70 independent promoter test (E. coli sigma70 promoter prediction) Dataset: E. coli sigma70 promoter dataset	0.9 Sensitivity fraction · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source ProkBERT-mini: E. coli sigma70 promoter prediction Test-only evaluation on Cassiano and Silva-Rocha 2020 data; methods have different training histories. 865 high-evidence RegulonDB 10.5 promoters and 1,000 nucleotide-distribution-matched negative sequences. Promoter exact matches removed from model training. Aggregation: Not reported ProkBERT family: genomic language models for microbiome applications · Table 3, row ProkBERT-mini, column Sensitivity; XML row2 column4
Configuration: ProkBERT-mini	Protocol: E. coli sigma70 independent promoter test (E. coli sigma70 promoter prediction) Dataset: E. coli sigma70 promoter dataset	0.85 Specificity fraction · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source ProkBERT-mini: E. coli sigma70 promoter prediction Test-only evaluation on Cassiano and Silva-Rocha 2020 data; methods have different training histories. 865 high-evidence RegulonDB 10.5 promoters and 1,000 nucleotide-distribution-matched negative sequences. Promoter exact matches removed from model training. Aggregation: Not reported ProkBERT family: genomic language models for microbiome applications · Table 3, row ProkBERT-mini, column Specificity; XML row2 column5
Configuration: ProkBERT-mini	Protocol: E. coli sigma70 independent promoter test (E. coli sigma70 promoter prediction) Dataset: E. coli sigma70 promoter dataset	0.74 MCC dimensionless · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source ProkBERT-mini: E. coli sigma70 promoter prediction Test-only evaluation on Cassiano and Silva-Rocha 2020 data; methods have different training histories. 865 high-evidence RegulonDB 10.5 promoters and 1,000 nucleotide-distribution-matched negative sequences. Promoter exact matches removed from model training. Aggregation: Not reported ProkBERT family: genomic language models for microbiome applications · Table 3, row ProkBERT-mini, column MCC; XML row2 column3

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

Use this model

How it works, versions and access

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

How it works

How the evaluated method works

A transformer encoder uses local-context-aware 6-mer tokenisation for the ProkBERT-mini variant, followed by self-supervised pretraining and task-specific fine-tuning.

Sources

What was evaluated

The linked evaluation record identifies ProkBERT-mini: E. coli sigma70 promoter prediction. Its dataset, split, adaptation and evidence origin remain attached to the reported results.

Sources

ProkBERT family: genomic language models for microbiome applications · The named evaluation’s methods and comparison table; exact preserved evaluation IDs: evaluation-lit-033

Strengths, limitations and unresolved questions

Strengths and limitations

Strengths and considerations

Targets microbial sequence distributions and evaluates promoter-related downstream tasks.
Sources
ProkBERT family: genomic language models for microbiome applications · 2 Materials and methods/2.3 Application I: bacterial promoter prediction/2.3.1 Dataset overview/2.3.1.2 Dataset construction for multispecies train, test and validation sets (paragraph 6); 3 Results and discussion/3.4 ProkBERT swiftly and accurately identifies phage sequences, even in challenging settings (paragraph 2)

Limitations and conditions

The mini checkpoint and tokenizer settings define the evaluated system; results cannot be transferred to every ProkBERT variant.
Sources
ProkBERT family: genomic language models for microbiome applications · 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 3 Results and discussion/3.3 ProkBERT performs accurately and robustly in promoter sequence recognition (paragraph 8)

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-4438513d9cd42c

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	Transformer representation pipeline; this record is the paper-specific evaluated configuration. Sources ProkBERT family: genomic language models for microbiome applications · 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 2 Materials and methods (paragraph 1)
Architecture / procedure	A transformer encoder uses local-context-aware 6-mer tokenisation for the ProkBERT-mini variant, followed by self-supervised pretraining and task-specific fine-tuning. Sources ProkBERT family: genomic language models for microbiome applications · 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 2 Materials and methods (paragraph 1)
Biological inputs	Microbial DNA sequences Sources ProkBERT family: genomic language models for microbiome applications · 1 Introduction (paragraph 7); 4 Conclusion (paragraph 10)
Outputs	DNA embeddings and task-specific genomic classifications Sources ProkBERT family: genomic language models for microbiome applications · 2 Materials and methods/2.5 Applied metrics (paragraph 1); 3 Results and discussion/3.1 ProkBERT's learned representations capture genomic structure and phylogeny (paragraph 4)
Parameters	Approximately 20 million parameters. Sources ProkBERT family: genomic language models for microbiome applications · 4 Conclusion (paragraph 8); 2 Materials and methods/2.2 Pretraining and learning sequence representations/2.2.2 Training process/2.2.2.2 Training phases and configuration (paragraph 1)
Known versions / configuration	ProkBERT-mini is the comparison-table label; that label does not specify an immutable weight revision. · Not reported in inspected sources Sources ProkBERT family: genomic language models for microbiome applications · Model identification in the comparison table and corresponding Methods; immutable checkpoint revision is not supplied by the table label.
Training data / fitting	Unlabelled microbial genome sequences followed by supervised task data, as specified in the paper. Sources ProkBERT family: genomic language models for microbiome applications · 2 Materials and methods/2.4 Application II: phage sequence analysis/2.4.2 Model training for phage sequence analysis (paragraph 1); 2 Materials and methods/2.2 Pretraining and learning sequence representations/2.2.4 Analysis of encoder outputs (paragraph 6)
Context limits	Approximately 1 kb for ProkBERT-mini; the distinct mini-long variant supports approximately 2 kb. Sources ProkBERT family: genomic language models for microbiome applications · 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); Table T3 (paragraph 1)
Access	Official study implementation and usage documentation: https://github.com/nbrg-ppcu/prokbert/blob/8670ae92b816cff158a0b85647a8dea122e251eb/README.md. This pinned documentation revision is not automatically the evaluated weight revision. Sources nbrg-ppcu/prokbert README.md · README.md; installation, model download and usage instructions
Code licence	MIT (study repository code at the cited revision; this does not establish every dependency or historical checkpoint licence). Sources nbrg-ppcu/prokbert 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 sources Sources nbrg-ppcu/prokbert README.md · README.md; checkpoint/access documentation and licence scope

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.

21 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
Diagram caption Conceptual input–method–output guide. Check the procedure text and linked evaluation for fitted components, additional inputs and exact settings. Individual claims	ProkBERT family: genomic language models for microbiome applications Original source ↗ 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 2 Materials and methods (paragraph 1) Version: PMC10810988.1 Retrieved: 2026-09-16T10:33:36.197Z	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: `8610e2a54aa877c8dc565a9cdb6e82099f284c5e0907a52cab18d994ea732436` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram steps Microbial DNA sequences ProkBERT-mini DNA embeddings and task-specific genomic classifications Individual claims	ProkBERT family: genomic language models for microbiome applications Original source ↗ 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 2 Materials and methods (paragraph 1) Version: PMC10810988.1 Retrieved: 2026-09-16T10:33:36.197Z	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: `8610e2a54aa877c8dc565a9cdb6e82099f284c5e0907a52cab18d994ea732436` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram title Evaluated procedure (conceptual) Individual claims	ProkBERT family: genomic language models for microbiome applications Original source ↗ 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 2 Materials and methods (paragraph 1) Version: PMC10810988.1 Retrieved: 2026-09-16T10:33:36.197Z	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: `8610e2a54aa877c8dc565a9cdb6e82099f284c5e0907a52cab18d994ea732436` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Model type Transformer representation pipeline; this record is the paper-specific evaluated configuration. Individual claims	ProkBERT family: genomic language models for microbiome applications Original source ↗ 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 2 Materials and methods (paragraph 1) Version: PMC10810988.1 Retrieved: 2026-09-16T10:33:36.197Z	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: `8610e2a54aa877c8dc565a9cdb6e82099f284c5e0907a52cab18d994ea732436` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Architecture / procedure A transformer encoder uses local-context-aware 6-mer tokenisation for the ProkBERT-mini variant, followed by self-supervised pretraining and task-specific fine-tuning. Individual claims	ProkBERT family: genomic language models for microbiome applications Original source ↗ 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 2 Materials and methods (paragraph 1) Version: PMC10810988.1 Retrieved: 2026-09-16T10:33:36.197Z	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: `8610e2a54aa877c8dc565a9cdb6e82099f284c5e0907a52cab18d994ea732436` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
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	nbrg-ppcu/prokbert README.md Original source ↗ README.md; checkpoint/access documentation and licence scope Version: 8670ae92b816cff158a0b85647a8dea122e251eb Retrieved: 2026-09-16T19:54:21.078483+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.10.value` Source artifact SHA-256: `29de39c6ad006ce704ab14240cfd97af93da411fb63ec89ebe499c2646928cfc` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Biological inputs Microbial DNA sequences Individual claims	ProkBERT family: genomic language models for microbiome applications Original source ↗ 1 Introduction (paragraph 7); 4 Conclusion (paragraph 10) Version: PMC10810988.1 Retrieved: 2026-09-16T10:33:36.197Z	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: `8610e2a54aa877c8dc565a9cdb6e82099f284c5e0907a52cab18d994ea732436` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Outputs DNA embeddings and task-specific genomic classifications Individual claims	ProkBERT family: genomic language models for microbiome applications Original source ↗ 2 Materials and methods/2.5 Applied metrics (paragraph 1); 3 Results and discussion/3.1 ProkBERT's learned representations capture genomic structure and phylogeny (paragraph 4) Version: PMC10810988.1 Retrieved: 2026-09-16T10:33:36.197Z	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: `8610e2a54aa877c8dc565a9cdb6e82099f284c5e0907a52cab18d994ea732436` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Parameters Approximately 20 million parameters. Individual claims	ProkBERT family: genomic language models for microbiome applications Original source ↗ 4 Conclusion (paragraph 8); 2 Materials and methods/2.2 Pretraining and learning sequence representations/2.2.2 Training process/2.2.2.2 Training phases and configuration (paragraph 1) Version: PMC10810988.1 Retrieved: 2026-09-16T10:33:36.197Z	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.4.value` Source artifact SHA-256: `8610e2a54aa877c8dc565a9cdb6e82099f284c5e0907a52cab18d994ea732436` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Known versions / configuration ProkBERT-mini is the comparison-table label; that label does not specify an immutable weight revision. Individual claims	ProkBERT family: genomic language models for microbiome applications Original source ↗ Model identification in the comparison table and corresponding Methods; immutable checkpoint revision is not supplied by the table label. Version: PMC10810988.1 Retrieved: 2026-09-16T10:33:36.197Z	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.5.value` Source artifact SHA-256: `8610e2a54aa877c8dc565a9cdb6e82099f284c5e0907a52cab18d994ea732436` 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

View linked audit checks and correction history

Release 2026-09-29-06401fd5b220 · Record review: needs review

3 source records and release history

ProkBERT family: genomic language models for microbiome applications · Original source · PMC10810988.1
nbrg-ppcu/prokbert README.md · Original source · 8670ae92b816cff158a0b85647a8dea122e251eb
nbrg-ppcu/prokbert LICENSE · Original source · 8670ae92b816cff158a0b85647a8dea122e251eb

Download this release

Technical metadata and extraction receipts

Stable ID: reported-model-4438513d9cd42c

areas: microbes-communities
entity level: method
version: Not reported
reported name: ProkBERT-mini
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: prokbert-2024; source locator: 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 2 Materials and methods (paragraph 1) | 2 Materials and methods/2.1 Sequence data/2.1.1 Sequence segmentation and tokenization (paragraph 6); 4 Conclusion (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.

Related records

family: ProkBERT
model: ProkBERT-mini: E. coli sigma70 promoter prediction
subject: ProkBERT-mini: family ProkBERT

Suggest a correction