AI ENABLEMENT

mRNA Data Generation

We generate large-scale mRNA datasets to train and validate AI models for optimized therapeutic design.

Data is a bottleneck for AI. Build your data moat.

Measurements that matter.

Not all measurements are relevant. Ginkgo’s in-cell assays for mRNA stability and translation efficiency are proven to support mRNA design for optimized protein expression in vivo and across different payloads.

Read our preprint

Validated screening platform.

Generating stability and translation data at scale is slow and resource-intensive. Accelerate discovery with tens of thousands of measurements in a validated assay - ready in as little as three months.

Get started

Purpose-built data.

Public datasets have limits. We measure what you design - train on diverse sequence libraries, validate predictions in the lab, and close the loop between AI-driven design and experimental validation. Explore our datasets or inquire about custom solutions.

Explore datasets

The Service

Get meaningful mRNA measurements for tens of thousands of your designs in a single experiment.

Fee-for-service. You own the data.

Two assay types that measure stability & translation efficiency

We construct pooled libraries of your 5' or 3' mRNA designs on plasmids, generate mRNA via in vitro transcription (IVT), transfect cells, and measure stability and translation efficiency using pooled assays with a quantitative NGS-based readout.

Download Sample Data

Stability Assays

mRNA stability over time up to 72 hours

Input

20,000 5’ or 3’ UTR sequences

After receiving your UTR sequences, we build and IVT an mRNA library of your designs.

Assay

In-cell mRNA stability assay

mRNA libraries are transfected into cells and RNA is harvested at four time points (3h, 24h, 48h, 72h). Each library member is measured across timepoints via NGS.

Output

Stability measurements per design

Stability scores are computed based on NGS counts. Processed datasets are returned to you, ready for model training and analysis. Six replicates across four timepoints lead to robust estimates.

Readout for mRNA data generation Stability assay

Trends across timepoints

Readout of an mRNA pooled stability assay. We produce thousands of data points for UTR candidates across timepoints. The y-axis represents the proportion of mRNA molecules that are remaining at each timepoint. An industry standard 3’UTR is shown at the bottom, and two synthetic Ginkgo 3’UTRs are shown above.

Download Sample Data

Translation Efficiency Assay

Relative ribosome load via polysome profiling

Input

20,000 5’ or 3’ UTR sequences

After receiving your UTR sequences, we build and IVT an mRNA library of your designs. mRNA libraries are transfected into cells and allowed to undergo translation.

Assay

In-cell translation efficiency assay

Cells are lysed, and ribosome-bound mRNAs are separated by sucrose gradient polysome fractionation. Each library member’s presence is measured across fractions via NGS.

Output

Translation efficiency measurements per design

Translation efficiency scores are computed based on NGS counts. Processed datasets are returned to you, ready for model training and analysis. Twelve fractions across multiple replicates lead to robust estimates.

Trends across polysome fractions

Readout of an mRNA pooled translation efficiency assay. Here we show each candidate’s presence per polysome fraction, which we then use to calculate an overall translation efficiency estimate.

Download Sample Data

Interact with Sample Datasets

Dataset Details	Stability-1	Translation-1
Release date	March 2025	March 2025
Data package	UTR stability estimate across timepoints UTR counts by sample and time point PDF outlining the experimental set up and sequence statistics	UTR translation efficiency estimate across fractions UTR counts by sample and fraction PDF outlining the experimental set up and sequence statistics
Sequence design	2625 genomic 3’ UTRs (subset of full experiment)	3189 genomic 5’ UTRs (subset of full experiment)
Readout	mRNA stability to 72 hours (3h, 24h, 48h, 72h)	Translation efficiency via polysome profiling
Assay context	HEK293T cells 5’ Cap1 150 nt poly-A tail Unmodified uridine GFP payload gene	HEK293T cells 5’ Cap1 120 nt poly-A tail N1-methylpseudouridine GFP payload gene
-	Download	Download

How we work with you

Step 01

Your sequence designs

Send Ginkgo your UTR sequence designs

Step 02

High-throughput pooled screens

Ginkgo executes library construction and high-throughput pooled screens

Step 03

Readouts

NGS-based readouts are captured for mRNA stability and/or translation efficiency

Step 04

Data transfer

Processed and QC'd measurements for each sequence are returned to you

We're improving RNA sequence design through AI

mDD-0: mRNA Discrete Diffusion for Generation of Stable mRNA Sequences

Engineering Stable mRNA: Insights from ML-Driven Design of 3' UTRs

Off-the-shelf AI-designed UTRs and custom solutions

Accelerate your genetic medicine development with licensable regulatory elements or more complex solutions. We offer off-the-shelf 5’ UTRs for increased translation and 3’ UTRs for increased stability, or we can develop a custom project to meet your RNA R&D needs.

License our 3' & 5' UTRs Explore more RNA solutions

Off-the-shelf licensable 3' and 5' UTR designs

Tell us what data you need. Our AI/ML experts can help you design the data generation campaign for your specific needs

Need data? Let's talk.

Schedule a call