AI in drug discovery and development: news watch

Recent developments in the biopharma AI world include a big financing for Verily, plus news from Zealand Pharma, Latent Labs, Insilico, and BMS/insitro.

Verily raises $300m, sheds Alphabet control

Verily has completed a private placement valued at $300 million, led by Series X Capital, that will be used to scale its AI-driven platform for clinical research, personalised care, and digital health technologies.

The financing also means that the Dallas, US company has transitioned away from majority control by Google parent Alphabet, although the tech company – which supported the financing alongside UCHealth and University of Colorado Anschutz – will retain a minority position. Verily will now operate as Verily Health, with greater leeway to forge new alliances as it endeavours to build digital infrastructure that can underpin the healthcare sector "from research to care."

Verily recently announced a number of commercial partnerships, including a new collaboration with Samsung's Galaxy Watch to help life sciences customers accelerate clinical research through biomarker development on its Pre platform.

Zealand sets up AI-focused R&D hub in Cambridge, MA

Denmark's Zealand Pharma is setting up a new research hub in Cambridge, Massachusetts, that will serve as its US headquarters and enhance its "research platform through AI-driven drug discovery, advanced automation, and next-generation molecule creation."

Zealand said it chose Cambridge for the hub as the location has a plethora of local partners that can help it to develop AI, machine learning, and automation models that "enhance the quality and efficiency of drug design and expand the pharmacological reach of the company's research platform" beyond its core peptide expertise.

Its objective is to add new treatment modalities like antibody-peptide conjugates (APCs) and small interfering RNA (siRNA) candidates that can complement its current focus on peptide candidates for metabolic diseases.

Latent Labs launches agent that designs antibodies from a text prompt

Generative AI (GenAI) specialist Latent Labs, founded in 2023 by former DeepMind researcher Dr Simon Kohl, has launched an AI agent that autonomously designs therapeutic antibodies from a text prompt, which the company claims can "compress…weeks of expert work into hours."

The agent is powered by Latent-X2, Latent Labs' recently launched frontier AI model that can design drug-like antibodies and peptides and reduce the need for wet lab work, and enables a single researcher to "run multiple design campaigns in parallel, across targets and modalities," according to Latent Labs.

The agent analyses the target molecules, applies biological reasoning to identify viable epitopes, designs antibody candidates using Latent-X2, validates them computationally, and iterates until the design goals are met.

Insilico Medicine rolls out natural language AI for drug researchers

Insilico Medicine has launched PandaClaw, a new autonomous AI agent that is an add-on to its PandaOmics drug discovery engine and automates complex analyses. According to the company, it can help researchers identify drug targets, uncover new indications, and build disease hypotheses using a natural language interface.

The emphasis on natural language is important, says Insilico, because conventional AI-enabled drug discovery can often require researchers to be 'bilingual' in biomedicine and AI, an effort that requires significant training. "PandaClaw lowers the barrier for biologists to apply AI, enabling sophisticated studies without specialised computational training," according to the Hong Kong and US-based company.

"Capable of executing complex tasks, bridging computational results with mechanistic interpretation, and delivering real-time results via simple natural language commands, PandaClaw represents the new standard in autonomous, accessible, and expert-level therapeutic discovery."

BMS broadens neurology collab with insitro

Six years after Bristol Myers Squibb and insitro formed a partnership to find new drug targets for neurodegenerative disorders, the alliance has been expanded to include two additional programmes in amyotrophic lateral sclerosis (ALS), prompting a $10 million payment to the South San Francisco biotech.

The two new ALS programmes (ALS-2 and ALS-3) add to another (ALS-1) that was nominated for further development by BMS at the end of 2024. insitro is working on an antisense oligonucleotide against the target, while simultaneously working on a small-molecule candidate for BMS.

All three have emerged from insitro's AI-driven Virtual Human platform, which applies machine learning, human genetics, and functional genomics to generate predictive in vitro models that provide insights into disease progression. The ALS programme is focusing on processes that modulate the effects of TDP-43 mislocalisation, which the partners believe is a central disease mechanism in nearly 97% of ALS patients.

Image by Gerd Altmann from Pixabay