The Value of ROH Metrics for Predicting Morbidity: Insights From a Large Cohort Analysis of Chromosomal Microarray.

Abstract

This retrospective cohort study aimed to define the optimal Regions of Homozygosity (ROH) size cut-offs for prediction of morbidity, based on 13 483 Chromosomal Microarray Analyses (CMA). Receiver operating characteristic (ROC) curves were generated, and area under the curve (AUC) was used to assess the predictive capability of total ROH percentage (TRPS), ROH number and ROH segment size in distinguishing between healthy (n=6,196) and affected (n=6,839) cohorts. The metrics were examined for telomeric and interstitial segments, distinct TRPS categories, and across different ancestral origins. ROH segments were identified in 13 035 samples (96.7%), encompassing 66 710 ROH segments. Significant differences in TRPS and ROH segment size were observed between healthy and affected cohorts (p=0.012 and p < 0.001, respectively). However, no clinically significant thresholds could be established based on ROC curves for TRPS and ROH number per sample, as well as for ROH size (AUC 0.64, 0.55, and 0.62, respectively, Figure 1). The same was noted for telomeric versus interstitial locations, various origins, and subcategories of TRPS. In conclusion, this study highlights the complexity of ROH interpretation and emphasizes the importance of tailored reporting strategies in clinical practice. Our findings underscore the need for context-specific reporting guidelines and further research, particularly in consanguineous populations.