Three new cohorts of Pakistani families segregating HL were examined by NGS. Cohorts 1, 2 and 3 have 261, 40, and 20 families, respectively, for a total of 321 families. For cohort 1, DNA samples from two affected individuals were first screened for variants in HGF (Schultz et al., 2009 (link)), CIB2 [c.272T>C; p.(Phe91Ser)] (Riazuddin et al., 2012 (link)), and GJB2 (Santos, Wajid, Pham, et al., 2005 (link)), using Sanger sequencing. HL-associated variants were found in 94 of these families. DNA samples from the family members of these 94 families were also Sanger-sequenced to test co-segregation of variants with HL. For the remaining 167 families that were negative for variants of GJB2, CIB2 or HGF, DNA samples for 129 families were genotyped using microsatellite or single nucleotide polymorphism (SNP) markers. Genotype data were checked for errors using PedCheck (O’Connell and Weeks, 1998) and MERLIN (Abecasis, et al. 2002). Analysis was performed using HomozygosityMapper (Seelow, Schuelke, Hildebrandt, & Nurnberg, 2009 (link)) and Allegro for linkage analysis (Gubjartsson, et al., 2005). Out of 129 families with genotype data, 46 families were mapped to a region containing a known HL gene and the putatively pathogenic variant was identified and co-segregation with HL confirmed using Sanger sequencing.
The remaining 121 families from cohort 1 included those who were not mapped to known HL genes and were negative when followed-up with Sanger sequencing. Forty additional families (cohort 2) did not have genome-scan genotype data. A DNA sample from an affected individual from each of these families (total 161 DNA samples) were exome sequenced either at the University of Maryland School of Medicine (UMSOM), University of Washington Center for Mendelian Genomics (UWCMG), or at the National Institute of Deafness and Other Communication Disorders (NIDCD) Genomic and Computational Biology Core.
Aside from the first two cohorts of 301 families, an additional 20 families (cohort 3) were screened using a gene panel that includes all known nonsyndromic and selected syndromic HL genes (n=101) (Zein et al., 2014 (link)) and the segregation of pathogenic variants identified from this panel was verified by Sanger sequencing the DNA of all informative family members.
Variants were considered further if (a) they have allele frequency <1% and (b) Combined Annotation Dependent Depletion (CADD; (Kircher et al., 2014 (link))) scaled score was ≥20 and/or (c) at least two of the following algorithms predict the mutation to be deleterious: PROVEAN (Y. Choi & Chan, 2015 (link)), SIFT (Kumar, Henikoff, & Ng, 2009 (link)), MutationTaster (Schwarz, Rodelsperger, Schuelke, & Seelow, 2010 (link)), MutationAssessor (Reva, Antipin, & Sander, 2007 (link)), FATHMM (Shihab et al., 2013 (link); Shihab et al., 2015 (link)) and PolyPhen2 (Adzhubei et al., 2010 (link)). Splice site and splice region variants were evaluated using splice site-altering scoring algorithms such as dbscSNV (Jian, Boerwinkle, & Liu, 2014 (link)) and Human Splice Finder (Desmet et al., 2009 (link)). Further details of bioinformatics analyses are described inSupp. Tables S1 and S2 . All variants were tested for co-segregation with HL in the corresponding family via Sanger sequencing. All novel variants have been submitted to ClinVar public databasis (https://www.ncbi.nlm.nih.gov/clinvar/ ).
The remaining 121 families from cohort 1 included those who were not mapped to known HL genes and were negative when followed-up with Sanger sequencing. Forty additional families (cohort 2) did not have genome-scan genotype data. A DNA sample from an affected individual from each of these families (total 161 DNA samples) were exome sequenced either at the University of Maryland School of Medicine (UMSOM), University of Washington Center for Mendelian Genomics (UWCMG), or at the National Institute of Deafness and Other Communication Disorders (NIDCD) Genomic and Computational Biology Core.
Aside from the first two cohorts of 301 families, an additional 20 families (cohort 3) were screened using a gene panel that includes all known nonsyndromic and selected syndromic HL genes (n=101) (Zein et al., 2014 (link)) and the segregation of pathogenic variants identified from this panel was verified by Sanger sequencing the DNA of all informative family members.
Variants were considered further if (a) they have allele frequency <1% and (b) Combined Annotation Dependent Depletion (CADD; (Kircher et al., 2014 (link))) scaled score was ≥20 and/or (c) at least two of the following algorithms predict the mutation to be deleterious: PROVEAN (Y. Choi & Chan, 2015 (link)), SIFT (Kumar, Henikoff, & Ng, 2009 (link)), MutationTaster (Schwarz, Rodelsperger, Schuelke, & Seelow, 2010 (link)), MutationAssessor (Reva, Antipin, & Sander, 2007 (link)), FATHMM (Shihab et al., 2013 (link); Shihab et al., 2015 (link)) and PolyPhen2 (Adzhubei et al., 2010 (link)). Splice site and splice region variants were evaluated using splice site-altering scoring algorithms such as dbscSNV (Jian, Boerwinkle, & Liu, 2014 (link)) and Human Splice Finder (Desmet et al., 2009 (link)). Further details of bioinformatics analyses are described in