Task

cross-species conservation prediction

Conservation prediction compares zero-shot sequence scores across genomic models with different context and prediction objectives.

Sources

PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table

1 evaluation · 1 metric row

At a glance

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.

Data, procedure and scoring
Property	Description and evidence
Datasets	Three conservation tasks use Sorghum bicolor with Andropogoneae alignments, a Poaceae coding-sequence alignment with Pharus latifolius as outgroup, and potato (Solanum tuberosum) with 95 Solanaceae genomes. Each uses its own alignment-derived conservation labels. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Evolutionary constraint prediction using the zero-shot strategy
Splits	This is zero-shot scoring rather than fitting on conservation labels. For the Andropogoneae and Poaceae tasks, the evaluated site is centered at position 4,096 in an 8,192-base window and its reference base is masked. Conservation thresholds and selected positions differ by task. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Evolutionary constraint prediction using the zero-shot strategy
Metrics	AUROC for conserved/non-conserved site classification in the cross-species conservation evaluations. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table
Baselines	PlantCAD2, PlantCAD, GPN and Evo2. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table
Leakage controls	The paper explicitly states that Solanaceae species were absent from PlantCAD2 pretraining, while Evo 2 included multiple Solanum genomes. This supports a scoped distinction for the potato experiment, not a universal claim that every evaluated position or homolog was unseen. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Results: PlantCAD2 accurately predicts evolutionary conservation with a zero-shot strategy; Methods: Evolutionary constraint prediction using the zero-shot strategy
Uncertainty	Supplemental Table 2 supplies task-, model- and context-specific AUROC point estimates, without uncertainty columns. Figure S2 varies context length; that variation is not a confidence interval or repeated-seed error estimate. · Not reported in inspected sources Sources (2) PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms; plantcad2-2025__media-1.xlsx · Supplemental Table 2 worksheet: Task, Model, Context and AUROC columns; Figure S2
Entity type	Paper-specific computational evaluation protocol. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table
Organisms	Plant genomic sequences. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table
Assays	Conservation/evolutionary-constraint annotations. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table
Allowed inputs	Plant DNA sequence. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table
Adaptation	Zero-shot scoring compared across PlantCAD2, PlantCAD, GPN and Evo2. Sources PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table

How it works

How it worksComputational evaluation flow

Conceptual summary of the cited evaluation; exact task configuration and source version remain part of the protocol.

Sources

Evaluation methodology

Three conservation tasks use Sorghum bicolor with Andropogoneae alignments, a Poaceae coding-sequence alignment with Pharus latifolius as outgroup, and potato (Solanum tuberosum) with 95 Solanaceae genomes. Each uses its own alignment-derived conservation labels. This is zero-shot scoring rather than fitting on conservation labels. For the Andropogoneae and Poaceae tasks, the evaluated site is centered at position 4,096 in an 8,192-base window and its reference base is masked. Conservation thresholds and selected positions differ by task. AUROC for conserved/non-conserved site classification in the cross-species conservation evaluations. PlantCAD2, PlantCAD, GPN and Evo2. The paper explicitly states that Solanaceae species were absent from PlantCAD2 pretraining, while Evo 2 included multiple Solanum genomes. This supports a scoped distinction for the potato experiment, not a universal claim that every evaluated position or homolog was unseen.

Sources

PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Evolutionary constraint prediction using the zero-shot strategy; Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table; Results: PlantCAD2 accurately predicts evolutionary conservation with a zero-shot strategy; Methods: Evolutionary constraint prediction using the zero-shot strategy

Recorded evaluations

Each evaluation records what was tested and under which conditions.

PlantCAD2: cross-species conservation prediction

Tested entities 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
PlantCAD2: cross-species conservation prediction Configuration: PlantCAD2Task: cross-species conservation predictionDataset: Andropogoneae genome-wide conservation Zero-shot score for conserved versus non-conserved sites from alignments of 35 Andropogoneae genomes Author-reported evaluation · Evaluation metadata: needs review
0.725 AUROC Unit: fraction · Direction: unknown	Uncertainty: not reported in legacy extract Scored: Not reported · Eligible: Not reported	source checkedPlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Table 1, Cross-species evolutionary conservation > Conservation within Andropogoneae (Genome-wide) row, PlantCAD2 AUROC entry Source checking is not independent reproduction.

Papers and result coverage

Last literature check: 2026-09-17. Primary-paper discovery and source inspection. Source-checked results are not independently reproduced experiments.

Paper or primary resource	Version	Reference
PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms	preprint version in PMC	Read source DOI: 10.1101/2025.08.27.672609

What is still missing

Table 1 is not a complete baseline comparison: named full-baseline scores are in Fig. 2/source workbook. Do not invent model identities from 'best benchmark'.
Keep lower PlantCAD2 non-TIS score as printed.

Search and extraction details

primary comparison table screened

Searches

"PMC12425018"

Evidence locations

Table 1
Figure 2
Results: cross-species conservation

Strengths and limitations

Strengths and considerations

No source-reviewed explanatory claims are recorded here yet.

Limitations and conditions

Short-context comparators receive cropped inputs. Evo2 uses a reference-allele likelihood because its autoregressive objective does not support identical masked-token scoring; those are material comparison differences.
Sources
PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table

Profile review details

Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change.

Stable record: reported-task-3109f8d0f2b7b5

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.

18 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 summary of the cited evaluation; exact task configuration and source version remain part of the protocol. Individual claims	PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms Original source ↗ Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table; Methods: Evolutionary constraint prediction using the zero-shot strategy Version: preprint version in PMC Retrieved: 2026-09-16T10:41:06Z	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.diagram.caption` Source artifact SHA-256: `4891955edb33af61e36dad51110574aa242e77c2df36429f100ff5a54bd8bd4b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram steps ["Input: Plant DNA sequence.","Evaluation: This is zero-shot scoring rather than fitting on conservation labels. For the Andropogoneae and Poaceae tasks, the evaluated site is centered at position 4,096 in an 8,192-base window and its reference base is masked. Conservation thresholds and selected positions differ by task.","Readout: AUROC for conserved/non-conserved site classification in the cross-species conservation evaluations."] Individual claims	PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms Original source ↗ Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table; Methods: Evolutionary constraint prediction using the zero-shot strategy Version: preprint version in PMC Retrieved: 2026-09-16T10:41:06Z	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.diagram.steps` Source artifact SHA-256: `4891955edb33af61e36dad51110574aa242e77c2df36429f100ff5a54bd8bd4b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram title Computational evaluation flow Individual claims	PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms Original source ↗ Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table; Methods: Evolutionary constraint prediction using the zero-shot strategy Version: preprint version in PMC Retrieved: 2026-09-16T10:41:06Z	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.diagram.title` Source artifact SHA-256: `4891955edb33af61e36dad51110574aa242e77c2df36429f100ff5a54bd8bd4b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Datasets Three conservation tasks use Sorghum bicolor with Andropogoneae alignments, a Poaceae coding-sequence alignment with Pharus latifolius as outgroup, and potato (Solanum tuberosum) with 95 Solanaceae genomes. Each uses its own alignment-derived conservation labels. Individual claims	PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms Original source ↗ Methods: Evolutionary constraint prediction using the zero-shot strategy Version: preprint version in PMC Retrieved: 2026-09-16T10:41:06Z	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.0.value` Source artifact SHA-256: `4891955edb33af61e36dad51110574aa242e77c2df36429f100ff5a54bd8bd4b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Splits This is zero-shot scoring rather than fitting on conservation labels. For the Andropogoneae and Poaceae tasks, the evaluated site is centered at position 4,096 in an 8,192-base window and its reference base is masked. Conservation thresholds and selected positions differ by task. Individual claims	PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms Original source ↗ Methods: Evolutionary constraint prediction using the zero-shot strategy Version: preprint version in PMC Retrieved: 2026-09-16T10:41:06Z	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.1.value` Source artifact SHA-256: `4891955edb33af61e36dad51110574aa242e77c2df36429f100ff5a54bd8bd4b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Adaptation Zero-shot scoring compared across PlantCAD2, PlantCAD, GPN and Evo2. Individual claims	PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms Original source ↗ Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table Version: preprint version in PMC Retrieved: 2026-09-16T10:41:06Z	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.10.value` Source artifact SHA-256: `4891955edb33af61e36dad51110574aa242e77c2df36429f100ff5a54bd8bd4b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Metrics AUROC for conserved/non-conserved site classification in the cross-species conservation evaluations. Individual claims	PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms Original source ↗ Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table Version: preprint version in PMC Retrieved: 2026-09-16T10:41:06Z	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.2.value` Source artifact SHA-256: `4891955edb33af61e36dad51110574aa242e77c2df36429f100ff5a54bd8bd4b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Baselines PlantCAD2, PlantCAD, GPN and Evo2. Individual claims	PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms Original source ↗ Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table Version: preprint version in PMC Retrieved: 2026-09-16T10:41:06Z	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.3.value` Source artifact SHA-256: `4891955edb33af61e36dad51110574aa242e77c2df36429f100ff5a54bd8bd4b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Leakage controls The paper explicitly states that Solanaceae species were absent from PlantCAD2 pretraining, while Evo 2 included multiple Solanum genomes. This supports a scoped distinction for the potato experiment, not a universal claim that every evaluated position or homolog was unseen. Individual claims	PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms Original source ↗ Results: PlantCAD2 accurately predicts evolutionary conservation with a zero-shot strategy; Methods: Evolutionary constraint prediction using the zero-shot strategy Version: preprint version in PMC Retrieved: 2026-09-16T10:41:06Z	source checked automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.4.value` Source artifact SHA-256: `4891955edb33af61e36dad51110574aa242e77c2df36429f100ff5a54bd8bd4b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Uncertainty Supplemental Table 2 supplies task-, model- and context-specific AUROC point estimates, without uncertainty columns. Figure S2 varies context length; that variation is not a confidence interval or repeated-seed error estimate. Individual claims	plantcad2-2025__media-1.xlsx Original source ↗ Supplemental Table 2 worksheet: Task, Model, Context and AUROC columns; Figure S2 Shared locator for this statement’s cited sources; not a separate locator for each citation. Version: Retrieved 2026-09-16; sha256:f551e80bf044a0ea8ccc2e6f443fdf3690ca5e4a54d0e4bb4b2e6470ffc944be Retrieved: 2026-09-16T21:05:57.965868+00:00	unreported automated source review · 2026-09-16 Audit details Targeted full-paper and supplement review of the outstanding task fields, with original dataset metadata checked where accessible. Source-scoped omissions are explicit; no independent benchmark reproduction or numerical-result change. Field: `attributes.profile.facts.5.value` Source artifact SHA-256: `f551e80bf044a0ea8ccc2e6f443fdf3690ca5e4a54d0e4bb4b2e6470ffc944be` Hash scope: Hash scope not separately documented; inspect source record Archive member: media-1.xlsx 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

PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Original source · preprint version in PMC
PlantCAD2: A Long-Context DNA Language Model for Cross-Species Functional Annotation in Angiosperms · Original source · preprint version in PMC
plantcad2-2025__media-1.xlsx · Original source · Retrieved 2026-09-16; sha256:f551e80bf044a0ea8ccc2e6f443fdf3690ca5e4a54d0e4bb4b2e6470ffc944be

Download this release

Technical metadata and extraction receipts

Stable ID: reported-task-3109f8d0f2b7b5

areas: dna-genomes
tasks: cross-species conservation prediction
entity level: task
version: Not reported
task: cross-species conservation prediction
scope note: Paper-specific evaluation task; protocol completeness requires further extraction.
benchmark research: review date: 2026-09-17; status: primary_comparison_table_screened; primary sources: expansion-p3-plantcad2-2025; inspected locators: Table 1; Figure 2; Results: cross-species conservation; searched queries: "PMC12425018"; gaps: Table 1 is not a complete baseline comparison: named full-baseline scores are in Fig. 2/source workbook. Do not invent model identities from 'best benchmark'.; Keep lower PlantCAD2 non-TIS score as printed.; claim scope: Primary-paper discovery and source inspection. Source-checked results are not independently reproduced experiments.
historical missing metadata: protocol version: not_reported_in_legacy_extract; split: 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: benchmark
entity classification: review date: 2026-09-17; rationale: This source-scoped record identifies the biological prediction task and holds its paper context. Preserve the existing task identity; exact split, model adaptation and scoring remain in linked evaluations or separate protocol records.; source ids: plantcad2-2025; source locator: Methods: Zero-shot evaluation of PlantCAD2, PlantCAD and GPN; Evo2 evaluation; cached text lines 110–115; task metric definitions and corresponding results table; ambiguities: A paper- or suite-specific task may constrain some inputs or metrics; that alone does not make it interchangeable with a complete versioned protocol. No protocol equivalence is inferred.; Some legacy profile Entity type facts use the generic phrase computational evaluation protocol. That boilerplate is not sufficient to establish a single fixed protocol identity or to merge this task with another protocol record.

Related records

dataset: Andropogoneae genome-wide conservation
benchmark: PlantCAD2: cross-species conservation prediction

Suggest a correction