The current project is a continuation of the previous project (NAISS 2023/5-396) aiming to complement on-going collaborative studies of multifactorial human OCD, using canine compulsive disorder (CCD) as a genetic model. Previous project PI Matteo Bianchi has ended his position in our research group but the project in ongoing. Here we will leverage our new dog reference genome (UU_Cfam_GSD_1.0) (PMID:33568770) and canine genetic panels (Dog10K) to analyse the genetics of compulsive disorders in many thousands of mix breed dogs. Specifically, we are currently working with whole genome data from dog samples from Darwin’s Ark resource (PMID:35482869): 3044 sequenced with low coverage (1x), 94 with high coverage (>30x) and 411 genotyped on the Axiom Canine Genotyping Array set A and B (ThermoFisher Scientific). These data has just been realigned to the new dog reference genome (UU_Cfam_GSD_1.0) (PMID:33568770) and imputed (using a panel of~2000 genomes from the Dog10K project data (http://www.dog10kgenomes.org/)) the low pass and the array data using GLIMPSE (PMID:33414550) and BEAGLE5 (doi:10.1016/j.ajhg.2018.07.015; beagle.22Jul22.46e.jar), respectively. With the resulting genetic data in place, we are now preforming genome-wide association studies (GWAS). Survey items (N=14) related to repetitive and compulsive behaviour were extracted from Darwin’s Ark Questionnaire (Morrill et al. 2022) and were used in an Exploratory factor analysis (EFA) including 12,989 dogs with answers on all questions. We used R-packages nFactors (v. 2.4.1.1) and base (v. 4.3.1) for evaluating number of factors to use, psych (v. 2.4.3) for EFA, mirt (v. 1.41) for calculating factor scores for individuals using item response theory. We are performing GWAS using Proportional Odds Logistic Mixed Model (POLMM) (https://doi.org/10.1016/j.ajhg.2021.03.019), as part of the GRAB R-package (https://wenjianbi.github.io/grab.github.io/), for the items describing specific CCD behaviours and GTCA’s Mixed Linear Model Association - Leave One Chromosome Out (MLMA-LOCO) [PubMed ID: 21167468] for CCD factors (continuous variables of items that correlate). Preliminary results show significant genetic associations with CCD, and CCD candidate genes implicate shared biological pathways with human OCD genes showing the potential for cross-species comparisons. This warrants further investigation of replication datasets of CCD from Sweden and fine-mapping of associated regions. In the fine-mapping step, we plan to use Zoonomia mammalian constraint scores (https://zoonomiaproject.org) to identify candidate functional variants. As a next step of this project we have already started a collection of phenotypes (via a newly designed and tailored behaviour survey) and genetic data (low-pass sequencing from saliva swabs) from a new cohort in Sweden/Europe which will serve as a replication cohort.