Table 1. Overview of LRP6 variants identified with MIP screen in OFC and TA patients vs. controls.

All variants were coding or canonical splice site variants with a population frequency of <0.1% (based on dbSNP142 and an in-house database with >5,000 samples), ≤3 samples in this study with the same variant and a ‘GATK quality by depth’ of >1,000, the latter was based on previous MIP data and extensive validations by Sanger sequencing showing low false positive rates and high sensitivity. Minimal average coverage over all MIPs of included samples was 100-fold. Most unique and rare non-synonymous variants reported here in cases have been validated by Sanger sequencing.

	Total cases w. variant (n=1139)	TA cases w. variant (n=67)	OFC cases w. variant (n=1072)	Controls w. variant (n=706)
Rare (<0.1%) coding & SS$ variants	30	8	22	16
Of which synonymous	3	1	2	3
Of which non-synonymous	27	7a	20b	13a,b
Unique/private variants	13	5	8	5
Of which synonymous	1	0	1	2
Of which non-synonymous	12	5c	7d	3c,d
Of which CADD PHRED >20	7	4e	3f	1e,f

^$Splice site canonical dinucleotide; CADD PHRED-like score (Kircher et al. 2014, Nature Genetics).

^{a, b}7 rare, non-synonymous variants in LRP6 in 67 cases with OD were highly significant when compared to the variant load in 706 controls (Fisher's exact test after Bonferroni correction P = 0.0056022); while 20 rare, non-synonymous variants in LRP6 in 1072 cases with OFC does not show significance when compared to the variant load in 706 controls.

^{c, d}5 unique, non-synonymous variants in LRP6 in 67 cases with OD were highly significant when compared to 3 such variants in 706 controls (Fisher's exact test after Bonferroni correction P = 0.0012354); while 7 unique, non-synonymous variants in LRP6 in 1072 cases with OFC does not show significance when compared to the variant load in 706 controls.

^{e, f}4 unique, predicted damaging variants in LRP6 in 67 cases with OD were highly significant when compared to 1 such variant in 706 controls (Fisher's exact test after Bonferroni correction P = 0.001515); while 7 unique, non-synonymous variants in LRP6 in 1073 cases with OFC does not show significance when compared to the variant load in 706 controls.