Bioinformatics Analysis of ABCA4 Variants in Terms of Pathogenicity Prediction

ABCA4 is a retina‐specific member of the ATP‐binding cassette transporter protein family. It is responsible for the transport and clearance of retinal derivatives produced in the visual cycle in the photoreceptor cells. More than a thousand mutations in the ABCA4 gene are associated with a series of autosomal recessive inherited retinal diseases, including Stargardt disease, retinitis pigmentosa, and cone‐rod dystrophy, and also linked with susceptibility for age‐related macular degeneration. Determining disease‐causing variants is a crucial part of clinical management, yet, a limited number of the ABCA4 variants have been studied in vitro and relatively few of them have been functionally characterized. Bioinformatics approaches can be highly effective to interpret the consequences of the mutations more rapidly. In this study, we aim to analyze the pathogenicity of ABCA4 clinical and prospective variants utilizing two methods to infer mutation effects: prediction software and computational protein models. We hypothesize that these two approaches in combination will allow us to identify the consequences of the ABCA4 variants and gain insight into the overall structure and function of the protein. We found that the pathogenicity prediction software results align well with the structural analysis results in most instances. The clinical in silico prediction tools were found to predict pathogenicity more frequently than structural analyses, albeit with high sensitivity and low specificity. On the other hand, structure prediction tools are high specificity and thus, predict more accurately. Our results also indicate a correlation between disease severity and structural changes in protein models induced by genetic variations. Our findings suggest that computational analyses are promising tools to predict pathogenicity and can aid in the understanding of disease‐associated ABCA4 genetic variants. These computational approaches have additional applications in protein analysis and provide a strong framework for future in vitro studies.