Configuration

mRNA-LM

mRNA-LM represents complete mRNAs by integrating separate models of the 5′ UTR, coding sequence and 3′ UTR.

Sources

mRNA-LM: full-length integrated SLM for mRNA analysis · Results/Evaluation of UTRBERTs (paragraph 1); Materials and methods/Datasets for pretraining UTRBERTs (paragraph 1)

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

mRNA-LM: full-length integrated SLM for mRNA analysis · Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/mRNA-LM: joint language model/Learning the joint representation of segment sequence using CLIP (paragraph 5)

At a glance

Model type

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

Sources

Biological inputs

The 5′ UTR, CDS and 3′ UTR of an mRNA

Sources

mRNA-LM: full-length integrated SLM for mRNA analysis · Conclusion (paragraph 14); Results/Evaluation of UTRBERTs (paragraph 1)

Outputs

Integrated mRNA representations and task-specific property predictions

Sources

mRNA-LM: full-length integrated SLM for mRNA analysis · Abstract (paragraph 1); Conclusion (paragraph 14)

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
mRNA-LM: mRNA half-life prediction Configuration: mRNA-LMTask: mRNA half-life predictionDataset: mRNA half-life average test performance across cross-validation splits Author-reported evaluation · Evaluation metadata: needs review
0.696 Spearman rho Unit: unitless · Direction: unknown	Uncertainty: not reported in legacy extract Scored: Not reported · Eligible: Not reported	source checkedmRNA-LM: full-length integrated SLM for mRNA analysis · Table 1, mRNA-LM row, mRNA half-life Spearman column Source checking is not independent reproduction.

How it works

How the evaluated method works

Two nucleotide-token UTRBERTs and a codon-token CodonBERT component are integrated using contrastive learning inspired by CLIP. Each UTR encoder has 12 layers, 12 heads and 768-dimensional hidden states.

Sources

What was evaluated

The linked evaluation record identifies mRNA-LM: mRNA half-life prediction. Its dataset, split, adaptation and evidence origin remain attached to the reported results.

Sources

mRNA-LM: full-length integrated SLM for mRNA analysis · The named evaluation’s methods and comparison table; exact preserved evaluation IDs: evaluation-b2-mrna-lm-2025

Strengths and limitations

Strengths and considerations

Models interactions among mRNA regions rather than using only a coding-sequence representation.
Sources
mRNA-LM: full-length integrated SLM for mRNA analysis · Introduction (paragraph 5); Results/Evaluation of UTRBERTs (paragraph 1)

Limitations and conditions

Component-specific tokenisation and context limits matter; property heads must be evaluated on their own assay and split.
Sources
mRNA-LM: full-length integrated SLM for mRNA analysis · Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); 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-54d974e8e08043

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	Learned representation pipeline; this record is the paper-specific evaluated configuration. Sources mRNA-LM: full-length integrated SLM for mRNA analysis · Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/mRNA-LM: joint language model/Learning the joint representation of segment sequence using CLIP (paragraph 5)
Architecture / procedure	Two nucleotide-token UTRBERTs and a codon-token CodonBERT component are integrated using contrastive learning inspired by CLIP. Each UTR encoder has 12 layers, 12 heads and 768-dimensional hidden states. Sources mRNA-LM: full-length integrated SLM for mRNA analysis · Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/mRNA-LM: joint language model/Learning the joint representation of segment sequence using CLIP (paragraph 5)
Biological inputs	The 5′ UTR, CDS and 3′ UTR of an mRNA Sources mRNA-LM: full-length integrated SLM for mRNA analysis · Conclusion (paragraph 14); Results/Evaluation of UTRBERTs (paragraph 1)
Outputs	Integrated mRNA representations and task-specific property predictions Sources mRNA-LM: full-length integrated SLM for mRNA analysis · Abstract (paragraph 1); Conclusion (paragraph 14)
Parameters	The integrated mRNA-LM model is reported to contain more than 260 million parameters; its UTR modules each use 12 layers, 12 heads and hidden width 768. Sources mRNA-LM: full-length integrated SLM for mRNA analysis · Materials and methods/Fine-tuning using LoRA (paragraph 1); Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1)
Known versions / configuration	mRNA-LM is the comparison-table label; that label does not specify an immutable weight revision. · Not reported in inspected sources Sources mRNA-LM: full-length integrated SLM for mRNA analysis · Model identification in the comparison table and corresponding Methods; immutable checkpoint revision is not supplied by the table label.
Training data / fitting	Millions of mRNA sequences from multiple species; UTR encoders use masked-language pretraining. Sources mRNA-LM: full-length integrated SLM for mRNA analysis · Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/Datasets for pretraining UTRBERTs (paragraph 1)
Context limits	5UTRBERT maximum length 512 tokens; 3UTRBERT maximum length 1,024 tokens. CDS uses a separate codon model. Sources mRNA-LM: full-length integrated SLM for mRNA analysis · Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Conclusion (paragraph 9)
Access	Official study implementation and usage documentation: https://github.com/Sanofi-Public/mRNA-LM/blob/d7538c9aadbceb59a8832904292b279d0a4c2d12/README.md. This pinned documentation revision is not automatically the evaluated weight revision. Sources Sanofi-Public/mRNA-LM README.md · README.md; installation, model download and usage instructions
Code licence	Sanofi academic/non-commercial licence; use outside those permissions requires the separate terms described in the file. (study repository code at the cited revision; this does not establish every dependency or historical checkpoint licence). Sources Sanofi-Public/mRNA-LM LICENSE.txt · LICENSE.txt; 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 Sanofi-Public/mRNA-LM 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

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	mRNA-LM: full-length integrated SLM for mRNA analysis Original source ↗ Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/mRNA-LM: joint language model/Learning the joint representation of segment sequence using CLIP (paragraph 5) Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558214+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: `3a23de3c672ec162d13561c483f180a73b550d717256deffdc9099accec205fd` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram steps ["The 5′ UTR, CDS and 3′ UTR of an mRNA","mRNA-LM","Integrated mRNA representations and task-specific property predictions"] Individual claims	mRNA-LM: full-length integrated SLM for mRNA analysis Original source ↗ Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/mRNA-LM: joint language model/Learning the joint representation of segment sequence using CLIP (paragraph 5) Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558214+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: `3a23de3c672ec162d13561c483f180a73b550d717256deffdc9099accec205fd` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Diagram title Evaluated procedure (conceptual) Individual claims	mRNA-LM: full-length integrated SLM for mRNA analysis Original source ↗ Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/mRNA-LM: joint language model/Learning the joint representation of segment sequence using CLIP (paragraph 5) Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558214+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: `3a23de3c672ec162d13561c483f180a73b550d717256deffdc9099accec205fd` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Model type Learned representation pipeline; this record is the paper-specific evaluated configuration. Individual claims	mRNA-LM: full-length integrated SLM for mRNA analysis Original source ↗ Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/mRNA-LM: joint language model/Learning the joint representation of segment sequence using CLIP (paragraph 5) Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558214+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: `3a23de3c672ec162d13561c483f180a73b550d717256deffdc9099accec205fd` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Architecture / procedure Two nucleotide-token UTRBERTs and a codon-token CodonBERT component are integrated using contrastive learning inspired by CLIP. Each UTR encoder has 12 layers, 12 heads and 768-dimensional hidden states. Individual claims	mRNA-LM: full-length integrated SLM for mRNA analysis Original source ↗ Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/mRNA-LM: joint language model/Learning the joint representation of segment sequence using CLIP (paragraph 5) Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558214+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: `3a23de3c672ec162d13561c483f180a73b550d717256deffdc9099accec205fd` 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	Sanofi-Public/mRNA-LM README.md Original source ↗ README.md; checkpoint/access documentation and licence scope Version: d7538c9aadbceb59a8832904292b279d0a4c2d12 Retrieved: 2026-09-16T19:54:18.716122+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: `896d2f64f8a236c9f99cad115feb97dab369bbf4c42bb6025df01bbae0bd740b` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Biological inputs The 5′ UTR, CDS and 3′ UTR of an mRNA Individual claims	mRNA-LM: full-length integrated SLM for mRNA analysis Original source ↗ Conclusion (paragraph 14); Results/Evaluation of UTRBERTs (paragraph 1) Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558214+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: `3a23de3c672ec162d13561c483f180a73b550d717256deffdc9099accec205fd` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Outputs Integrated mRNA representations and task-specific property predictions Individual claims	mRNA-LM: full-length integrated SLM for mRNA analysis Original source ↗ Abstract (paragraph 1); Conclusion (paragraph 14) Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558214+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: `3a23de3c672ec162d13561c483f180a73b550d717256deffdc9099accec205fd` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Parameters The integrated mRNA-LM model is reported to contain more than 260 million parameters; its UTR modules each use 12 layers, 12 heads and hidden width 768. Individual claims	mRNA-LM: full-length integrated SLM for mRNA analysis Original source ↗ Materials and methods/Fine-tuning using LoRA (paragraph 1); Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1) Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558214+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.4.value` Source artifact SHA-256: `3a23de3c672ec162d13561c483f180a73b550d717256deffdc9099accec205fd` Hash scope: Hash scope not separately documented; inspect source record Inspected artifact
Known versions / configuration mRNA-LM is the comparison-table label; that label does not specify an immutable weight revision. Individual claims	mRNA-LM: full-length integrated SLM for mRNA analysis Original source ↗ Model identification in the comparison table and corresponding Methods; immutable checkpoint revision is not supplied by the table label. Version: journal full text in PMC Retrieved: 2026-09-16T10:38:57.558214+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.5.value` Source artifact SHA-256: `3a23de3c672ec162d13561c483f180a73b550d717256deffdc9099accec205fd` 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

mRNA-LM: full-length integrated SLM for mRNA analysis · Original source · journal full text in PMC
Sanofi-Public/mRNA-LM README.md · Original source · d7538c9aadbceb59a8832904292b279d0a4c2d12
Sanofi-Public/mRNA-LM LICENSE.txt · Original source · d7538c9aadbceb59a8832904292b279d0a4c2d12

Download this release

Technical metadata and extraction receipts

Stable ID: reported-model-54d974e8e08043

areas: rna-transcriptomes
entity level: method
version: not stated in table
reported name: mRNA-LM
historical missing metadata: 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: mrna-lm-2025; source locator: Materials and methods/Datasets for pretraining UTRBERTs/Pretraining our models (paragraph 1); Materials and methods/mRNA-LM: joint language model/Learning the joint representation of segment sequence using CLIP (paragraph 5) | Results/Evaluation of UTRBERTs (paragraph 1); Materials and methods/Datasets for pretraining UTRBERTs (paragraph 1); 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: mRNA-LM: mRNA half-life prediction

Suggest a correction