The study of rare diseases is important not only for the individuals affected but also for the advancement of medical knowledge and a deeper understanding of human biology and genetics. The wide repertoire of structural information now available from reliable and accurate prediction methods provides the opportunity to investigate the molecular origins of most of the rare diseases reviewed in the Orpha.net database. Thus, it has been possible to analyze the topology of the pathogenic missense variants found in the 2515 proteins involved in Mendelian rare diseases (MRDs), which form the database for our structural bioinformatics study. The amino acid substitutions responsible for MRDs showed different mutation site distributions at different three-dimensional protein depths. We then highlighted the depth-dependent effects of pathogenic variants for the 20,061 pathogenic variants that are present in our database. The results of this structural bioinformatics investigation are relevant, as they provide additional clues to mitigate the damage caused by MRD.

OBJECTIVE: Cancer risks conferred by germline, heterozygous, ATM pathogenic/likely pathogenic variants (PSVs) are yet to be consistently determined. The current study assessed these risks by analysis of a large dataset of ATM heterozygote loss of function (LOF) and missense PSV carriers tested with a multigene panel (MGP).
METHODS: De-identified data of all individuals who underwent ATM sequencing as part of MGP between October 2015 and February 2020 were reviewed. In cancer cases, rates for the six most prevalent variants and for all LOF and missense PSV combined were compared with rates of the same PSV in ethnically matched, healthy population controls. Statistical analysis included Chi-square tests and odds ratios calculations.
RESULTS: For female breast cancer cases, LOF )1794/219,269) and missense (301/219,269) ATM PSVs were seen at higher rates compared to gnomAD non-cancer controls (n = 157/56,001 and n = 27/61,208; p < 0.00001, respectively). Notably, the rate of the c.103C > T variant was higher in controls than in breast cancer cases [p = 0.001; OR 0.31 (95% CI 0.1-0.6)]. For all cancer cases combined, compared with non-cancer population controls, LOF (n = 143) and missense (n = 15) PSVs reported in both datasets were significantly more prevalent in cancer cases [ORLOF 1.7 (95% 1.5-1.9) ORmissense 3.0 (95% CI 2.3-4); p = 0.0001].
CONCLUSIONS: Both LOF and missense heterozygous ATM PSVs are more frequently detected in cases of several cancer types (breast, ovarian, prostate, lung, pancreatic) compared with healthy population controls. However, not all ATM PSVs confer an increased cancer risk (e.g., breast).