Model type
Dilated convolutional genomic-track predictor
Configuration
Basenji2
Basenji predicts quantitative regulatory activity along DNA and scores the effects of sequence changes.
Sources (2)
calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successorHow it worksBasenji workflow
Conceptual summary of the documented data flow; optional inputs and configured downstream stages must be reported for a reproducible evaluation.
Sources (2)
calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successorOverview
Inputs
DNA sequence and, for training, aligned quantitative regulatory measurements.
Sources (2)
calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successorOutputs
Predicted regulatory signal across sequence bins and derived variant-effect scores.
Sources (2)
calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successorAccess
Official project documentation and implementation: https://github.com/calico/basenji
Sources (2)
calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successorlimited source coverage · Automated source review, 2026-09-16. All specifications and missing details
Evaluations and results
1 evaluation · 1 metric rows. Different protocols are not a single leaderboard.
Filter evaluations
Applied filters: All linked evaluations
| Tested configuration | Protocol and dataset | Finding | Evidence and details |
|---|---|---|---|
| Configuration: Basenji2 | Protocol: Enformer CAGE gene-expression comparison Across genes CAGE Pearson: Mean across-experiment Pearson correlation of human test-gene CAGE expression Dataset subset: Enformer/Basenji2 human held-out test genes and CAGE experiments (Enformer CAGE gene-expression comparison split) | 0.81 mean_across_experiment_gene_pearson dimensionless · higher Uncertainty: Not reported Coverage: Not reported scored / Not reported eligible | Author-reported evaluation · source checkedMethods, coverage and sourceHuman held-out test-set protein-coding genes; CAGE read counts summed over all unique TSS locations, using TSS-overlapping 128-bp bin plus two neighboring bins, log(1+x)-transformed and standardized across genes separately per experiment. Pearson correlation across genes per CAGE experiment, then mean across experiments. Same Basenji2 dataset and genomic intervals; human split 34,021 training/2,213 validation/1,937 test sequences. Cross-species homologous 1 Mb regions partitioned by connected components. Test-time mean of 8 random ≤3 bp shift/reverse-complement augmentations. Enformer input 196,608 bp, Basenji2 131,072 bp; receptive fields differ. Validation used for tuning; main comparison on test set. Aggregation: Not reported Enformer primary article: exact Figure 1b across-genes comparison reported in text · Results / Enformer improves gene expression prediction, Par7; Fig1b left caption; Methods Par32–36; exact narrative comparison, Basenji2 value 0.81 |
Source checking is not independent reproduction. Release 2026-09-23-2b89723c6dd9.
Use this model
How it works, versions and access
Related profile: Basenji. This page retains the exact record and its evaluation context.
This configuration
Pretrained Basenji2 used in the main Enformer paper comparisons (Methods Par35); a configured Basenji family version, not original Basenji1.
- record
- Basenji2
- configuration
- Not reported
- entity type
- Configuration
How it works
How it works
Basenji predicts quantitative regulatory activity along DNA and scores the effects of sequence changes. Deep convolutional sequence model with binned regression outputs. The documented inputs are DNA sequence and, for training, aligned quantitative regulatory measurements. The output consists of predicted regulatory signal across sequence bins and derived variant-effect scores.
Sources (2)
calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successorVersions and reproducibility
Basenji implementation; Basset, Akita and Saluki are described separately. The applicable input limits require configuration-specific checking.
Sources (2)
calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successorStrengths, limitations and unresolved questions
Strengths and limitations
Strengths and considerations
- The same framework supports quantitative signal prediction and nucleotide/variant attribution workflows.
Sources (2)
calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor
Limitations and conditions
- The repository also contains Akita and Saluki, which are separate models. Their inputs and tasks must not be assigned to Basenji simply because they share a repository.
Sources (2)
calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor
Profile review details
Inspected pinned official documentation, relevant implementation files and named primary-paper sections. Claims are limited to those artifacts. Remaining field extraction and identity conflicts are explicit; no new performance claims, model runs or human review are implied.
Stable record: discovery-model-basenjiSpecifications
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 | Dilated convolutional genomic-track predictorSources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Architecture | Deep convolutional sequence model with binned regression outputs.Sources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Inputs | DNA sequence and, for training, aligned quantitative regulatory measurements.Sources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Outputs | Predicted regulatory signal across sequence bins and derived variant-effect scores.Sources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Parameters | The repository describes a configurable Basenji model family rather than one checkpoint with a common parameter count; the selected model configuration is required. · Not reported in inspected sourcesSources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Known versions | Basenji implementation; Basset, Akita and Saluki are described separately.Sources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Training data | Chosen regulatory-activity datasets; the README points to preprocessing/training tutorials rather than identifying one universal checkpoint.Sources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Training cutoff | No checkpoint is selected by this family record. Training-track accessions and collection dates must be taken from the chosen Basenji release, not inferred from the repository date. · Not reported in inspected sourcesSources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Context limits | The inspected family README does not fix one input and output window across Basenji configurations. A run must preserve its sequence length, pooling and output-bin settings. · Not reported in inspected sourcesSources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Weights licence | Separate checkpoint-distribution terms are not stated in the inspected release documentation and licence material. The source-code licence alone is not recorded as an explicit weight grant. · Not reported in inspected sourcesSources (3)calico/basenji: README.md; calico/basenji: docs/train.md; calico/basenji: LICENSE · README.md: Basenji and Basset successor; LICENSE: licence text |
| Access | Official project documentation and implementation: https://github.com/calico/basenjiSources (2)calico/basenji: README.md; calico/basenji: docs/train.md · README.md: Basenji and Basset successor |
| Code licence | Apache-2.0Sourcescalico/basenji: LICENSE · LICENSE: licence text |
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.
1 evidence row matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Relationship: family discovery-model-basenji Individual claims | Enformer primary article: exact Figure 1b across-genes comparison reported in text Results / Enformer improves gene expression prediction, Par7; Fig1b left caption; Methods Par32–36 Version: PMC8490152; SHA-256 pinned XML snapshot | source checked automated source review · 2026-09-23 Audit detailsSource-backed evaluated identity only; no independent reproduction. Field: Claim: enformer-cage-2021-method-basenji2-discovery-model-basenji-identity-claim Source artifact SHA-256: Hash scope: Hash scope not separately documented; inspect source record |
Sources and history
View linked audit checks and correction history
Release 2026-09-23-2b89723c6dd9 · Record review: source checked
1 source records and release history
- Enformer primary article: exact Figure 1b across-genes comparison reported in text · Original source · PMC8490152; SHA-256 pinned XML snapshot
Technical metadata and extraction receipts
Stable ID: enformer-cage-2021-method-basenji2
- areas
- genomics
- source locator
- Results / Enformer improves gene expression prediction, Par7; Fig1b left caption; Methods Par32–36
- missing metadata
- checkpoint revision: unreported; parameters: unextracted