Background: Autism Spectrum Disorder (ASD) forms a phenotypically and genetically heterogeneous group of neurodevelopmental conditions which are highly heritable. Being able to identify more confined subgroups of this condition may enable identification of different sets of genetic pathways involved and inform a more accurate clinical classification. In my PhD project, I aim to identify genetically-informed subgroups of autism patients through genome wide association studies (GWAS) of a range of autism behaviours, followed by patient similarity network of their polygenic risk scores. For this task, I use Simon Simplex Collection data for over 2500 individuals, rich in phenotypic information from across autism questionnaires. I performed factor analysis of ~300 individual questions and constructed 10 latent factors representing main domains of autism behaviours. To gain more granular insight into the conditions, instead of performing GWAS of a large umbrella term of autism diagnosis, I will perform GWAS of each of such defined behaviours. I will use the results to calculate polygenic scores of each behaviour and further utilise those to construct patient similarity network. The hope is that this will reveal clusters of patients with similar genetic background and will translate to clinically relevant subgroups of patients.