Describing familial aggregation of infectious disease is an important first step in identifying genetic components of disease resistance. Traditional measures of familial aggregation utilize subjects of known pedigree in family-based designs to compare disease frequency between family members of cases and family members of controls. Restricting the study-base to populations with known pedigree structures in paratuberculosis research may introduce selection bias if study populations are not representative in regards to other risk factors for disease. We developed an alternative approach for identifying groups of genetically similar individuals to compare disease frequency. Texas beef cattle herds were selected for sampling based on breeder surveys, veterinarian referral, and clinical cases admitted to the Texas Veterinary Medical Center. Diagnostic samples were collected from animals (n = 2,622) >2 years of age and evaluated for paratuberculosis using 2 ELISAs (Herdchek^®, IDEXX Laboratories; Parachek^®, Prionics) and fecal culture. Additionally, whole blood was preserved for genotyping. All animals positive on at least 1 test and 3 herd-matched controls were selected for genotyping. A separate group of animals of known pedigree relationships were sampled for genotyping to assess validity of clustering methods. Cases, controls, and pedigreed animals were genotyped using a panel of 12 microsatellites previously described for parentage testing in cattle. Bayesian methods (Structure v2.2) were employed to identify the optimal number of clusters of genetically similar individuals in the sample population and to probabilistically assign individuals to clusters. The proportion of parent-offspring pairs assigned to the same cluster among the pedigreed animals was evaluated to validate clustering methods. Conditional logistic regression was used to compare proportion of test positive animals among clusters controlling for herd of residence. Analysis of cluster results indicated that 9 clusters was optimal for this population. Using the cluster with the lowest proportion of test-positive animals as the referent group, the odds of having a positive test were significantly greater for 3 clusters (Cluster 2 OR 36.4, 95% CI 3.1 to 430.4; Cluster 7 OR 7.4, 95% CI 1.0 to 12.0; Cluster 9 OR 5.9, 95% CI 1.8 to 19.4) compared to the referent cluster. Of the 9 animals positive for Mycobacterium avium subsp. paratuberculosis on fecal culture, 5 were assigned to cluster 7. These results support the hypothesis that there are differences in genetic susceptibility to paratuberculosis test-positivity that can be quantified for beef cattle of unknown pedigree using genetic markers to assemble groups of genetically similar individuals. Employed methods demonstrate a unique approach to describing familial aggregation of disease in cattle. Clusters with the disparate odds can be targeted for further study of genetic susceptibility and may contribute to the understanding of the pathogenesis and genetic resistance to paratuberculosis in cattle.