Deep Apple Therapeutics launches with small-molecule drug discovery program

Deep Apple Therapeutics has launched with $52 million in Series A funding from Apple Tree Partners to develop small-molecule drugs for metabolic disorders. Its founders include Stanford University structural biologist Georgios Skiniotis, and University of California, San Francisco (UCSF), computational biologists Brian Shoichet and John Irwin.

One of the largest families of human membrane proteins are G-protein coupled receptors (GPCRs), which are involved in cellular signaling for a variety of physiological functions, including metabolism, smell, and sight. GPCRs are also one of the most drugged protein families, and interest in GPCRs as a therapeutic target has increased in recent years. One high-profile class of GPCR targeting drugs are the glucagon-like peptide-1 agonists that the US Food and Drug Administration has approved for weight loss, including Novo Nordisk’s Wegovy in June 2021 and Eli Lilly’s Zepbound in November 2023.

Deep Apple will use its founders’ expertise in structural and computational biology to uncover novel GPCR targeting small molecules for treating metabolic, immunology, and inflammation disorders. The company will use ensemble cryogenic electron microscopy (cryo-EM) to identify different active GPCR conformations and reveal unique pockets in the receptors that can be exploited as therapeutic targets.

Once these novel binding pockets are identified, Deep Apple will conduct in silico screening of its large virtual libraries to identify potential drugs. “We believe virtual screening as a technology has made great progress, but is still relatively underrepresented in the success stories of drug discovery,” says Spiros Liras, Deep Apple’s CEO, citing a study in theJournal of Medicinal Chemistry that less than 1% of recently discovered clinical candidates came from a virtual screen (2023, DOI: 10.1021/acs.jmedchem.3c00521).

Deep Apple’s deep learning methodologies will translate 2D chemotypes into 3D molecular shapes to aid the design of suitable drug candidates, Liras says. The company is also using deep learning to create a compound scoring algorithm for the results of its virtual screening process.

Despite its use of novel AI and deep learning technologies, Liras notes that it isn’t the company’s sole asset. “Deep learning is an enabler of accelerated drug discovery—but we don’t want to label Deep Apple as ‘an AI company,’” Liras says.

The company, based in the San Francisco Bay area, will use its start-up funds to advance multiple programs through the preclinical stage. Its portfolio includes an inflammation-related target and several weight loss and management targets currently in lead optimization.