Leveraging the power of new molecular technologies in the clinical setting requires unprecedented awareness of limitations and drawbacks: experience of one diagnostic laboratory.

Abstract

Advances in molecular technologies and in-silico variant prediction tools offer wide-ranging opportunities in diagnostic settings, yet they also present with significant limitations. Here, we contextualise the limitations of next-generation sequencing (NGS), multiplex ligation-dependent probe amplification (MLPA) and in-silico prediction tools routinely used by diagnostic laboratories by reviewing specific experiences from our diagnostic laboratory. We investigated discordant annotations and/or incorrect variant 'callings' in exons of 56 genes constituting our cardiomyopathy and connective tissue disorder NGS panels. Discordant variants and segmental duplications (SD) were queried using the National Center for Biotechnology Information (NCBI) Basic Local Alignment Search Tool and the University of California Santa Cruz genome browser, respectively, to identify regions of high homology. Discrepant variant analyses by in-silico models were re-evaluated using updated file entries. We observed a 5% error rate in <i>MYH7</i> variant 'calling' using MLPA, which resulted from >90% homology of the <i>MYH7</i> probe-binding site to <i>MYH6</i>. SDs were detected in <i>TTN</i>, <i>PKP2</i> and <i>MYLK</i>. SDs in <i>MYLK</i> presented the highest risk (15.7%) of incorrect variant 'calling'. The inaccurate 'callings' and discrepant in-silico predictions were resolved following detailed investigation into the source of error. Recognising the limitations described here may help avoid incorrect diagnoses and leverage the power of new molecular technologies in diagnostic settings.