FIGURES

Figure 1 | AutoCaSc workflow and categories

a) Schematic of the proposed workflow embedding AutoCaSc into a genetic pipeline for identification of pathogenic variants in individuals with NDD. When trio exome sequencing does not lead to a diagnosis, AutoCaSc is intended to prioritize variants and thus focus on the most promising candidates. b) Composition of the CaSc with the four categories, as well as the maximum achievable points: Variant attributes (max. 6 points), Gene plausibility (max. 6 points), Inheritance (max. 3 points), Gene constraint (max. 1 point). The maximum number of points for inheritance is awarded for autosomal recessive variants where the variant occurs in at least one other affected sibling. Gene constraint metrics, however, are only awarded for autosomal dominant inherited variants, which is why both categories together score a maximum of 3 points. c) Depiction of the AutoCaSc tools front-ends. Schematic illustration of vcfAutoCaSc, which is intended for when whole VCFs or cohorts are to be screened (left side). Screenshot of the results page of webAutoCaSc, which is intended to evaluate single or small groups of variants quickly and without installation (right side).

Figure 2 | Simulated and real trio scoring

a) Achieved CaSc of the inserted variants, subdivided according to the affected novel NDD gene. The color of the circles encodes the mode of inheritance. The red dashed line represents a possible cutoff of 6. b) ROC curve relating the false positive rate and true positive rate for the simulation experiment. optimal cutoff at CaSc 6 marked in red. c) Scoring results for the 78/93 (83.9%) in-house trio exomes with at >0 candidate variants scored. In the left part of the figure, the trios are plotted on the x-axis in descending order of the highest candidate score (CaSc), the y-axis shows the CaSc. The dots represent candidate variants. The large dots are the variants with the highest score in a trio. The colors indicate whether a candidate was scored only manually (n=2), only by AutoCaSc (n=230), or by both (n=79). The right part shows a histogram of the achieved CaSc.