To tackle all three, we treated the search as one optimization problem at XtalPi. Generative models proposed structures, predictive models assessed them, and robots synthesized candidates for screening. The results then shaped which molecules we designed next.

The >1,000:1 figure compares exposure in gut tissue with exposure in blood. In preclinical studies, KQTD-126 also inhibited TRKA, TRKB, and TRKC at IC₅₀ <1 nM and showed high selectivity over other kinases.

How can AI models and chemistry robots help design a pan-TRK inhibitor that stays in the gut? We just submitted a US IND for KQTD-126, XtalPi’s internally developed, potential first-in-class candidate for chronic intestinal pain, designed through an AI and robotics loop. Preclinically, gut-to-blood exposure exceeds 1,000:1.

The >1,000:1 figure compares drug exposure in gut tissue with exposure in blood. In preclinical studies, KQTD-126 also inhibited TRKA, TRKB, and TRKC at IC₅₀ <1 nM and showed high selectivity over other kinases.

Most #HIV drugs stop the virus after it is already inside the cell. A small molecule that stops it at the door is a different bet, and a harder one. Viva-Thera, which we incubated, has nominated a preclinical candidate against that entry machinery in 7 months. The target: MPER

Around the #MPER biology, Viva-Thera is also building gene-delivery and cell-therapy approaches — four modalities pointed at prevention, treatment, and the reservoir that existing regimens leave behind.