Background Diff erentiating type 1 diabetes from other subtypes in ethnic groups is challenging. Non-white young adults are assumed to be more likely to develop type 2 diabetes, although the characteristics of type 1 diabetes have not been studied in non-white ethnic groups. Data suggest that second generation migrant populations adopt the local risk for type 1 diabetes, but the phenotype in these groups remains unexplored. We aimed to investigate the phenotype of type 1 diabetes in a UK multi-ethnic population. Methods The MY DIABETES study is a multicentre cross-sectional study systematically phenotyping people diagnosed with any type of diabetes under 30 years of age from white or South Asian ancestry. In this preliminary analysis we report the clinical and biochemical characteristics, including islet-cell antibody status (GAD and IA-2), of people with type 1 diabetes (fasting C-peptide <201 pmol/L) by ethnicity and generation of migration. The study continues to recruit about 50 participants per month across 25 sites. Findings Of 427 recruits, 305 (71%) were white and 76 (18%) South Asian. Other ethnic groups included African-Caribbean, middle-eastern, and south east Asian. Criteria for type 1 diabetes were met in 218 white (87%) and 34 South Asian (54%) participants (p<0•0001). South Asian and white participants had similar ages at diagnosis (median 21•6 years vs 16•2), body-mass index (26•0 kg/m² vs 26•5), and glycated haemoglobin (62 mmol/mol for both). No diff erence in the proportion with detectable antibodies was observed (56% [122/217] white vs 56% [19/34] South Asian), when adjusted for duration. Islet-cell antibody status was similar (GAD positive 102 [47%] of 218 white vs 16 [48%] of 34 South Asian; IA-2 positive 21 [9•6%] vs 3 [9•7%]). Second generation South Asian people (UK born) were diagnosed at a younger age than were fi rst generation people (median age 11•4 years [IQR 3•1-24•4] vs 23•3 [15•7-27•1], p=0•039) but had similar durations of disease (median 22•8 years [IQR 10•9-36•8] vs 19•8 [6•6-27•3]), and antibody positivity did not diff er between the groups (11/16 [68•0%] vs 7/15 [46•7%]). Interpretation Young onset type 1 diabetes in South Asian and white people in the UK is phenotypically very similar. These data suggest that in young people presenting with diabetes, ethnicity should not have an impact on clinical diagnosis, even in those without detectable antibodies. Migration generation might aff ect phenotype, and additional investigations are needed to study diff erences between native and migrant ethnic groups. Work is ongoing to further characterise type 1 diabetes and other subtypes in the MY DIABETES cohort.