Validating the Sapientia inheritance filters to facilitate implementing trio clinical exome sequencing into diagnostic service

Data analysis is failing to keep pace with the volume of data generated by next generation sequencing (NGS) technology; a variant filtering strategy is required to streamline NGS analysis. Trio-based inheritance filtering offers the advantage of focusing interpretation on clinically-relevant de-novo, recessive or X-linked variants. SapientiaTM is a clinical decision support software. SapientiaTM inheritance filters can filter trio-NGS data based on the inheritance (e.g. maternal, de-novo, biparental) or mode of inheritance (e.g. autosomal dominant, X-linked recessive). When the inheritance is unclear, it is important to determine a logical order of using the inheritance filters. A series of experiments spiking-in variants of known inheritance were used to compare the number of genes in each inheritance filtration output and to assess the efficiency of each tested inheritance filter to identify the spike-in variants correctly. To streamline trio analysis, this study developed an order for sequential inheritance filtering in the context of unaffected parents (average number of genes filtered): X-linked recessive (if male) (1), autosomal recessive homozygous (2), autosomal recessive compound heterozygous (7) and de-novo (11). The rationale is to progress in ascending order of filtered number of genes to accelerate analysis. Specific inheritance filters were disregarded from the analysis workflow if they a) filtered clinically-insignificant pedigrees, b) increased the number of filters requiring user operation or c) were not functioning as intended. This study has demonstrated that, apart from the X-linked dominant filter, the SapientiaTM inheritance filters perform as intended. Analysis of the inheritance-filtered variant output is manageable within the resources of a diagnostic laboratory.