Genomics Collaboration Senior Scientist Genome-Wide Homozygosity Mapping and Linkage Analysis in Consanguineous Families Using the Affymetrix Mapping 10K Array Identifies a Novel Deafness Locus Carsten Rosenow, Ph.D. Genomics Collaboration Senior Scientist Genotyping

Systems Biology Workflow Mapping 10K Array Expression Array Identify candidate regions Linkage Fine Mapping Differentially Expressed Genes Clinical analysis Functional analysis CustomSeq

Genetic Analysis Workflow Analysis Tool Experimental Approach Mapping10K Linkage LOH HBD Fine Mapping Two Affymetrix products can be used for Genetic analysis. First the Mapping 10K array for applications like Linkage, LOH and HBD (Homozygosity by descent) and second the CustomSeq array for sequence analysis and novel SNP discovery. Candidate regions identified by either Expression analysis or Mapping 10K analysis can be used mapped using the CustomeSeq platform Sequence Analysis Novel SNP discovery CustomSeq™

GeneChip Mapping 10K

Presentation overview Technical Overview Application: Homozygosity by descent analysis Linkage Mapping Data Analysis Spotfire decision site

PCR: One Primer Amplification GeneChip® Mapping 10K Assay Overview 250 ng Genomic DNA Xba RE Digestion Adaptor Ligation PCR: One Primer Amplification Complexity Reduction The new developed assay for the mapping 10K array is the cornerstone of the product. Traditionally genotyping was conducted using locus specific PCR. Since this approach is not scalable we have developed a novel assay tgo map more than 10.000 SNPs using a complexity reduction and amplification schema. The Mapping assay begins with only 250 ng genomic DNA. It is digested with a restriction enzyme, such as XbaI and we are going to get a whole range of sizes We ligate a common set of adaptors, which will ligate to all the size fragments. Next, a PCR step is optimized to only amplify fragments in the 250-1000 bp range (as shown in red in the first fragment). Followed by fragmentation, labeling, and hybridization on the same platform used for expression analysis. We have selected our SNPs from the TSC database based on their presence in this “Xba” fraction of the genome. (see next few) Fragmentation and Labeling AA BB AB Hyb & Wash

GeneChip Mapping 10K Tiling Strategy High probe redundancy 40 probes (PM and MM) per SNP 20 probes per SNP allele 10 probes per DNA strand Forward and reverse strand interrogated An important feature of our accuracy is probe redundancy. Historically we interrogate every gene with currently 11 probes which gives us a good estimate of gene expression. Accuracy in Genotyping is an important requirement since genotype errors can lead to inflated or deflated lod scores. We use 40 probes for each SNP interrogated. Each position has a perfect match and mismatch probe. 20 probes per allele or 10 per strand. We interrogate forward and reverse strand.

GeneChip Mapping 10K tiling strategy Allele 1 C G Forward Reverse SNP position T A -3 -1 +2 +4 Allele 1 C G Forward Reverse Allele 2 A T In addition to the center or SNP position we interrogate neighboring bases (from -4 to +4) with probes consisting of perfect match and mismatch probes.

Genome Coverage: 400 Microsatellite Markers Dec 2002 Golden Path Dec 2002 Golden Path Blue = STR from CIDR panel Black = Gaps Red = at least 1 SNP per 100 kb Black = Gaps Median intermarker distance: 4.7 Mb Mean intermarker distance: 5.6 Mb Mean genetic gap distance: 8.9 cM Average Heterozygosity 0 .76 Here is the coverage of the 400 marker set used by CIDR. Note every chromosome has markers, but it is somewhat uneven. Black represent GAPs in the human genome, where there is no sequence. Median intermarker distance: 105 kb Mean intermarker distance: 210 kb Mean genetic gap distance: 0.32 cM Average Heterozygosity 0 .37

Linkage analysis Information content: Measurement of the goodness of a set of markers Function of marker heterozygosity and number of meiosis in a study For multipoint linkage analysis marker density is also important Nuclear families

Information content comparison Affymetrix Mapping 10K - Marshfield

SNP Annotation NetAffx™ Analysis Center

Homozygosity by Descent study (HBD) Molecular Otolaryngology Research Laboratories, University of Iowa Richard JH Smith Nicole Meyer

Homozygosity by descent (HBD) mapping for rare autosomal recessive disorders Basic principle of homozygosity by descent. Each parent gives one chromosome to an offspring. In an inbreeding family the ancestor allele (here in yellow) gets inherited by the affected child since both parents are carrier of the allele. If this allele contains a recessive SNP the child would be affected (recessive homozygote). In addition to linkage analysis the children can be analysed for large regions of homozygosity by descent for the identification of the disease causing allele. HBD

Homozygosity mapping for rare disorders In rare autosomal recessive diseases, inbred pedigrees are highly informative When parents are consanguineous, affected children are homozygous for alleles identical by descent (IBD) One affected child of a 1st cousin marriage is as informative as 3 affected children of unrelated parents Linkage mapping can be carried out by searching for regions of increased marker homozygosity Single pedigree can be powerful enough to map rare recessive traits

Study Consanguineous family studied Four out of five children are affected with deafness Hereditary deafness affects 1:2,000 newborns and accounts for greater than 50% of severe-to-profound childhood deafness

Workflow Type >10K genotypes in the family parents, affected and unaffected children Mendel inheritance error check replace Mendel error with nocall Homozygosity by descent (HBD) high stringency analysis treat no calls as homozygous regions look for homozygous regions in all affected individuals Disregard regions which are also homozygous in parents and in the unaffected children Identify ancestor haplotypes

Pedigree Autosomal recessive disease

Analysis Analysis done using 400 microsattelite markers HBD region identified on Chromosome 1 Analysis done with GeneChip Mapping 10K array HBD region on Chromosome 1 confirmed Novel HBD locus identified on Chromosome 3 Lod scores for both regions is 2.8

Chromosome 1: HBD region confirmed PARENTS 3 4 5 6 7 1 SNP1 2 MS1 SNP2 MS2 MS3 MS4 MS5 SNP3 SNP4 SNP5 SNP6 MS6 MS7 SNP7 SNP8 SNP9 SNP10 SNP11 SNP12 SNP13 SNP14

Chromosome 3 Homozygosity by descent

Homozygosity by Descent Analysis in Spotfire Decision Site

Import raw data Import raw data from database, text files, or clipboard Easily join patient phenotype information

Utilize annotations from NetAffx http://www. affymetrix NetAffx provides annotation files with known SNP attributes (e.g. chromosome, location, etc…)

Annotations enable visualization Each blue line represents one SNP

Perform HBD calculations (1) To start, count instances of homozygosity among affected individuals. Marked records below show two SNPs where 4 of 4 individuals are homozygous for this allele.

Perform HBD calculations (2) Generate statistical significance of homozygosity over “sliding” window of user-defined size along chromosomes.

Visualize results (1) Visualize chromosome map using color and size to reflect regions of significant homozygosity.

Visualize results (2) Or use slider to focus on most relevant SNPs, mark, and add column as reminder.

Visualize results (3) Visualize the significance over the chromosome using the line chart, and drill down to specific locations.

Incorporate known genetic markers Add known disease markers from OMIM and visualize alongside SNPs (genes as dark blue marks)

Are relevant genes near significant SNPs? Step 1: Focus on area of significance on chromosome 3 (zoom using range sliders) (zoom range)

Are relevant genes near significant SNPs? Step 2: Filter down to relevant genes (blue x’s) using text search query device Filter Closest known deafness locus ~15 x 106 bp Conclusion: New deafness loci is yet to be discovered in this region

Incorporate metrics from other software By manually parsing and importing the results from MERLIN linkage software, one can visualize linkage results. (same significant region on chr 3)

Acknowledgements Molecular Otolaryngology Research Laboratories, University of Iowa Richard JH Smith Nicole Meyer Affymetrix Shoulian Dong Rui Mei Spotfire Brendan Gibson Kristen Stoops