Mosaic Biosciences Launches SortAI, a Wet Lab Validation Service Built for AI-Designed Antibody Libraries
Published: May 25, 2026 | Category: News, Technology
Mosaic Biosciences today announced the commercial launch of SortAI, a yeast display-based validation service designed specifically for AI-generated antibody sequence panels. SortAI enables AI drug discovery companies to screen model-derived libraries before committing to full recombinant protein production, returning structured experimental data that feeds directly back into model retraining.
The Bottleneck AI Drug Discovery Can’t Ignore
Generative AI platforms are transforming antibody drug design, producing campaigns of hundreds of thousands to millions of computationally designed sequences per cycle. Yet the infrastructure built to validate those sequences was designed around traditional single-candidate drug programs: full recombinant expression on every sequence, one at a time.
The result is a compounding problem: slow cycles, wasted production budget, and models that cannot improve until the wet lab catches up. Typical CROs produce and characterize whatever sequence list arrives in their inbox. SortAI filters it first.
“Generative models can produce thousands of antibody candidates. Wet labs can validate a fraction of them. SortAI triages AI-generated libraries at scale so your best sequences get tested first and nothing falls through.”
— Eric Furfine, PhD, Chief Scientific Officer, Mosaic Biosciences
One Engagement. The Full Triage Picture.
SortAI delivers a complete triage picture in a single defined engagement:
Binding, affinity, epitope binning, developability, and function together: A complete discovery funnel in a single engagement, with deliverables structured for model retraining rather than drug program documentation.
Sequence liability review at intake: Mosaic scientists screen panels for CDR liabilities, aggregation motifs, poor expression signals, and other flags most AI algorithms don’t account for. Bad sequences are removed before they waste wet lab capacity.
Yeast-based selection at library scale: Multi-round selection across panels ranging from hundreds to over millions of sequences, with next-generation sequencing (NGS) at each round to track full population dynamics.
Negative data returned as a deliverable: Every selection round generates a negative population. SortAI sequences and returns those alongside ranked positive binders — because what didn’t bind is half the training signal.
Function pulled forward: Candidates that pass binding are immediately pressure-tested through cell-based functional screening, before expensive downstream development work begins.
Binding, affinity, epitope binning, developability, and function together: A complete discovery funnel in a single engagement, with deliverables structured for model retraining rather than drug program documentation.
Faster Cycles. A Fraction of the Cost.
A production-centric competitor workflow covering binding and affinity alone requires approximately 7 weeks and costs significantly more than a full SortAI engagement, which extends through developability profiling and cell-based functional screening. Sequence enrichment for model training is available as early as 6 weeks from project initiation; a full engagement completes in approximately 14 weeks.
The yeast display platform underpinning SortAI scales to 10⁷ sequence diversity, three orders of magnitude beyond the throughput of production-centric validation workflows, enabling functional triage at the true scale of AI-generated libraries.
“AI companies need data that makes the next design cycle better, not just data that’s fast. SortAI is built on the principle that quality of signal matters more than speed of throughput, and that the two aren’t mutually exclusive.”
— Tracey Mullen, Chief Strategy Officer, Mosaic Biosciences
Built on Demonstrated Capability
SortAI is backed by Mosaic’s team of 50+ scientists, 80% holding advanced degrees and more than 216 completed discovery programs. The yeast display workflow that powers SortAI has already demonstrated results in library-based screening engagements that competing CROs were unable to accommodate.
The platform was first introduced publicly at PEGS Boston 2026 and is now available as a defined commercial service. Detailed pricing and scope information is available upon request under NDA.
Learn more at mosaicbio.com/sortai
Read the full press release here.
About Mosaic Biosciences
Mosaic Biosciences is a biologics discovery company specializing in antibody and protein engineering, with expertise spanning transgenic mouse platforms, display technologies, and multispecific antibody formats. Mosaic partners with biotech and pharmaceutical companies to generate high-quality, developable biologics optimized for real-world therapeutic success.
To learn more about our discovery capabilities or explore a partnership, contact us.
Read the full press release:
Speak to a scientist
Connect with our discovery experts to discuss your program and explore how Mosaic can help move your ideas forward.