Latent Labs offers Latent‑X, an atom‑resolution generative AI model that designs de novo protein binders—specifically macrocycles and mini‑binders—directly from a target structure. Its no‑code web platform generates and ranks hundreds of candidate sequences and conformations in minutes, delivering designs with experimentally validated hit rates up to 100 % and picomolar affinities, thus reducing synthesis and testing cycles for biotech and pharmaceutical R&D.
Funding
$40M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Traditional protein binder discovery relies on high-throughput screening of millions of random molecules, yielding hit rates well below 1% and requiring months of wet‑lab work and large budgets. This low efficiency hampers rapid development of therapeutic antibodies, enzymes, and other biologics.
Solution
Latent Labs offers Latent‑X, an atom‑level generative AI model that designs de novo protein binders—specifically macrocycles and mini‑binders—directly from a target structure. The model co‑samples amino‑acid sequences and three‑dimensional conformations, enforcing atomic‑level biochemical constraints such as hydrogen‑bonding and π‑stacking. Users access the technology through a no‑code web platform that handles target upload, hotspot definition, binder generation, and computational ranking, delivering designs in seconds. Laboratory validation demonstrates 91‑100 % hit rates for macrocycles and 10‑64 % for mini‑binders, with affinities reaching the picomolar range, dramatically reducing the number of candidates that must be synthesized and tested.
Target Audience
Primary customers are biotech and pharmaceutical R&D teams, protein‑engineering groups, and academic laboratories that need rapid, high‑confidence design of protein binders for therapeutic or research applications.
Features
- Atom‑resolution generative model that simultaneously predicts protein sequence and structure, obeying physical chemistry rules.
- Supports multiple therapeutic binder modalities, currently macrocycles (cyclic peptides) and protein mini‑binders, with plans for nanobodies and antibodies.
- Generates designs 10× faster than prior methods; typical batches of 100 candidates are produced within minutes.
- Laboratory‑validated hit rates of 91‑100 % for macrocycles and up to 64 % for mini‑binders, achieving single‑digit micromolar to picomolar binding affinities.
- Integrated computational filters (structure prediction, confidence metrics) rank designs and assign “Pass” scores to prioritize lab‑ready candidates.
- No‑code web interface enables target upload, hotspot selection, binder generation, and visualization without requiring AI expertise or infrastructure.
- Free tier with daily user credits; commercial licensing grants non‑exclusive rights to generated sequences.