We propose to develop designs for phase I trials conducted in heterogeneous groups of patients. Phase I trials in oncology are meant to establish the 'maximally tolerated dose' (MTD), the highest dose that can be administered without excessive side effects. In many studies, the group structure is not used in the design. The resulting recommended dose is weighted in favor of the dose for the most frequently occurring group, effectively moving away from 'personalized' dosing for patients. At present, the most common way to include the group structure as part of the design is to conduct a separate phase I trial within each group. This can be inefficient, requiring a large number of patients, and can lead to dose recommendations that run counter to what is known clinically about the groups. In this proposal, we will develop efficient designs for phase I trials in groups that build from our previous work in designs for combinations of agent. Our goal is to develop efficient and more accurate designs for phase I trials in groups of patients that can lead to improved care across the entire range of cancers and cancer therapies.