Inversions are known contributors to the pathogenesis of genetic diseases. Identifying inversions poses significant challenges, making it one of the most demanding structural variants (SVs) to detect and interpret. Recent advancements in sequencing technologies and the development of publicly available SV datasets have substantially enhanced our capability to explore inversions. However, a cross-comparison in those datasets remains unexplored. In this study, we reported a proband with familial hemophagocytic lymphohistiocytosis type-3 carrying a splicing variant (c.1389+1G>
A) in trans with an inversion present in 0.006345% of individuals in gnomAD (v4.0) that disrupts UNC13D. Based on this result, we investigate the features of potentially pathogenic inversions in gnomAD which revealed 98.9% of them are rare and disrupt 5% of protein-coding genes associated with a phenotype in OMIM. We then conducted a comparative analysis of additional public datasets, including DGV, 1KGP, and two recent studies from the Human Genome Structural Variation Consortium which revealed common and dataset-specific inversion characteristics suggesting methodology detection biases. Next, we investigated the genetic features of inversions disrupting the protein-coding genes. Notably, we found that the majority of protein-coding genes in OMIM disrupted by inversions are associated with autosomal recessive phenotypes supporting the hypothesis that inversions in trans with other variants are potential hidden causes of monogenic diseases. This effort aims to fill the gap in our understanding of the molecular characteristics of inversions with low frequency in the population and highlight the importance of identifying them in rare disease studies.