Craig Teerlink

REVEL: An Ensemble Method for Predicting the Pathogenicity of Rare Missense Variants

The vast majority of coding variants are rare, and assessment of the contribution of rare variants to complex traits is hampered by low statistical power and limited functional data. Improved methods for predicting the pathogenicity of rare coding variants are needed to facilitate the discovery of disease variants from exome sequencing studies. We developed REVEL (rare exome variant ensemble learner), an ensemble method for predicting the pathogenicity of missense variants on the basis of individual tools: MutPred, FATHMM, VEST, PolyPhen, SIFT, PROVEAN, MutationAssessor, MutationTaster, LRT, GERP, SiPhy, phyloP, and phastCons. REVEL was trained with recently discovered pathogenic and rare neutral missense variants, excluding those previously used to train its constituent tools. When applied to two independent test sets, REVEL had the best overall performance (p < 10-12) as compared to any individual tool and seven ensemble methods: MetaSVM, MetaLR, KGGSeq, Condel, CADD, DANN, and Eigen. Importantly, REVEL also had the best performance for distinguishing pathogenic from rare neutral variants with allele frequencies <0.5%. The area under the receiver operating characteristic curve (AUC) for REVEL was 0.046-0.182 higher in an independent test set of 935 recent SwissVar disease variants and 123,935 putatively neutral exome sequencing variants and 0.027-0.143 higher in an independent test set of 1,953 pathogenic and 2,406 benign variants recently reported in ClinVar than the AUCs for other ensemble methods. We provide pre-computed REVEL scores for all possible human missense variants to facilitate the identification of pathogenic variants in the sea of rare variants discovered as sequencing studies expand in scale.

Evidence for an inherited predisposition contributing to the risk for rotator cuff disease.

BACKGROUND A genetic predisposition has been suggested to contribute to the risk for development of rotator cuff disease on the basis of observed family clusters of close relatives. We used a population-based resource combining genealogical data for Utah with clinical diagnosis data from a large Utah hospital to test the hypothesis of excess familial clustering for rotator cuff disease. METHODS The Utah Population Database contains combined health and genealogical data on over two million Utah residents. Current Procedural Terminology, Fourth Revision, codes (29827, 23412, 23410, and 23420) and International Classification of Diseases, Ninth Revision, codes (726.1, 727.61, and 840.4) entered in patient records were used to identify patients with rotator cuff disease. We tested the hypothesis of excess familial clustering using two well-established methods (the Genealogical Index of Familiality test and the estimation of relative risks in relatives) in the overall study group (3091 patients) and a subgroup of the study group diagnosed before the age of forty years (652 patients). RESULTS The Genealogical Index of Familiality test in patients diagnosed before the age of forty years showed significant excess relatedness for individuals with rotator cuff disease in close and distant relationships (as distant as third cousins) (p = 0.001). The relative risk of rotator cuff disease in the relatives of patients diagnosed before the age of forty years was significantly elevated for second degree (relative risk = 3.66, p = 0.0076) and third degree (relative risk = 1.81, p = 0.0479) relatives. CONCLUSIONS We analyzed a set of patients with diagnosed rotator cuff disease and a known genealogy to describe the familial clustering of affected individuals. The observations of significant excess relatedness of patients and the significantly elevated risks to both close and distant relatives of patients strongly support a heritable predisposition to rotator cuff disease.

Significant Linkage Evidence for a Predisposition Gene for Pelvic Floor Disorders on Chromosome 9q21

Predisposition factors for pelvic floor disorders (PFDs), including pelvic organ prolapse (POP), stress urinary incontinence (SUI), urge urinary incontinence (UUI), and hernias, are not well understood. We assessed linkage evidence for PFDs in mostly sister pairs who received treatment for moderate-to-severe POP. We genotyped 70 affected women of European descent from 32 eligible families with at least two affected cases by using the Illumina 1 million single-nucleotide polymorphism (SNP) marker set. Parametric linkage analysis with general dominant and recessive models was performed by the Markov chain Monte Carlo linkage analysis method, MCLINK, and a set of SNPs was formed, from which those in high linkage disequilibrium were eliminated. Significant genome-wide evidence for linkage was identified on chromosome 9q21 with a HLOD score of 3.41 under a recessive model. Seventeen pedigrees (53%) had at least nominal evidence for linkage on a by-pedigree basis at this region. These results provide evidence for a predisposition gene for PFDs on chromosome 9q.

A comprehensive survey of cancer risks in extended families

Purpose:Cancer is familial; yet known cancer predisposition genes, as well as recognized environmental factors, explain only a small percentage of familial cancer clusters. This population-based description of cancer clustering describes patterns of cancer coaggregation suggestive of a genetic predisposition.Methods:Using a computerized genealogy of Utah families linked to a statewide cancer registry, we estimated the relative risks for 36 different cancer sites in first-, second-, and third-degree relatives of cancer cases, for each cancer site individually, and between cancer sites. We estimated the sex- and birth-year-specific rates for cancer using 1 million individuals in the resource. We applied these rates to groups of cases or relatives and compared the observed and expected numbers of cancers to estimate relative risks.Results:Many cancer sites show significantly elevated relative risks among distant relatives for cancer of the same site, strongly supporting a heritable contribution. Multiple combinations of cancer sites were observed among first-, second-, and third-degree relatives, suggesting the existence of heritable syndromes involving more than one cancer site.Conclusion:This complete description of coaggregation of cancer by site in a well-defined population provides a set of observations supporting heritable cancer predispositions and may support the existence of genetic factors for many different cancers.Genet Med 2012:14(1):107–114

Significant evidence for a heritable contribution to cancer predisposition: a review of cancer familiality by site

Background/AimsSound and rigorous well-established, and newly extended, methods for genetic epidemiological analysis were used to analyze population evidence for genetic contributions to risk for numerous common cancer sites in Utah. The Utah Population Database (UPDB) has provided important illumination of the familial contribution to cancer risk by cancer site.MethodsWith over 15 years of new cancer data since the previous comprehensive familial cancer analysis, we tested for excess familial clustering using an expanded Genealogical Index of Familiality (dGIF) methodology that provides for a more informative, but conservative test for the existence of a genetic contribution to familial relatedness in cancer.ResultsSome new cancer sites have been analyzed for the first time, having achieved sufficiently large sample size with additions to the UPDB. This new analysis has identified 6 cancer sites with significant evidence for a heritable contribution to risk, including lip, chronic lymphocytic leukemia, thyroid, lung, prostate, and melanoma.ConclusionsBoth environmentally and genetically-based familial clustering have clinical significance, and these results support increased surveillance for cancer of the same sites among close relatives of affected individuals for many more cancers than are typically considered.

Prostate cancer risk prediction based on complete prostate cancer family history.

Prostate cancer (PC) relative risks (RRs) are typically estimated based on status of close relatives or presence of any affected relatives. This study provides RR estimates using extensive and specific PC family history.

Significant association of full-thickness rotator cuff tears and estrogen-related receptor-β (ESRRB)

BACKGROUND The precise etiology of rotator cuff disease is unknown, but prior evidence suggests a role for genetic factors. Variants of estrogen-related receptor-β (ESRRB) have been previously associated with rotator cuff disease. The purpose of the present study was to confirm the association between multiple candidate genes, including ESRRB, and rotator cuff disease in an independent set of patients with rotator cuff tear. MATERIALS AND METHODS The Illumina 5M (Illumina Inc, San Diego, CA, USA) single nucleotide polymorphism (SNP) platform was used to genotype 175 patients with rotator cuff tear. Genotypes were used to select a set of 2595 genetically matched Caucasian controls available from the Illumina iControls database. Tests of association were performed with Genome-wide Efficient Mixed Model Association (GEMMA) software at 69 SNPs that fell within 20 kb of 6 candidate genes (DEFB1, DENND2C, ESRRB, FGF3, FGF10, and FGFR1). RESULTS Tests of association revealed 1 significantly associated SNP occurring in ESRRB (rs17583842; P = 4.4E-4). Another SNP within ESRRB (rs7157192) had a nominal P value of 7.8E-3. FastPHASE software estimated 2 frequent haplotypes among 54 individuals who carried both risk alleles at these 2 SNPs. The first haplotype had a frequency of 13.9% (n = 15) in risk-allele carriers and only 2.2% in controls (odds ratio, 6.9; 95% confidence interval, 3.9-2.2). The second haplotype had a frequency of 12.9% in risk-allele carriers and only 2.7% in controls (odds ratio, 5.3; 95% confidence interval, 3.0-9.5). CONCLUSIONS The significant association and the presence of high-risk haplotypes identified in the ESRRB gene confirm the association of variants in ESRRB and rotator cuff disease.

Analysis of high-density single-nucleotide polymorphism data: three novel methods that control for linkage disequilibrium between markers in a linkage analysis

We performed a multipoint linkage analysis for rheumatoid arthritis (RA) using high-density single-nucleotide polymorphism (SNP) data for chromosome 6 and chromosome 21 using Genetic Analysis Workshop 15 (GAW15) data. These regions were previously shown to have high LOD scores, not accounting for linkage disequilibrium (LD). We propose three novel methods to control for LD in a linkage analysis: allow for LD between markers using graphical modeling, eliminate high-LD markers by principal-component analysis (PCA) using haplotype data, and eliminate high-LD markers by PCA using genotype data. All three novel methods were compared to the previously published SNPLINK high-LD elimination method. Although all four methods verified the previous results, differences in linkage peak height and position were observed across methods. Additional work is required to further understand the effects of LD on linkage results and explore LD control methodology.

Identification of a genetic variant associated with rotator cuff repair healing

BACKGROUND A familial and genetic predisposition for the development of rotator cuff tearing has been identified. The purpose of this study was to determine if a familial predisposition exists for healing after rotator cuff repair and if the reported significant association with a single-nucleotide polymorphism (SNP) in the ESRRB gene is present in patients who fail to heal. MATERIALS AND METHODS The study recruited 72 patients undergoing arthroscopic rotator cuff repair for a full-thickness posterosuperior tear. Magnetic resonance imaging studies were performed at a minimum of 1 year postoperatively (average, 2.6 years). Healing failures were classified as lateral or medial. Self-reported family history of rotator cuff tearing data and genome-wide genotypes were available. Characteristics of cases with and without a family history of rotator cuff tearing were compared, and a comparison of the frequency of SNP 1758384 (in ESRRB) was performed between patients who healed and those who failed to heal. RESULTS Of the rotator cuff repairs, 42% failed to heal; 42% of patients reported a family history of rotator cuff tear. Multivariate regression analysis showed a significant association between familiality and overall healing failure (medial and lateral failures) (P = .036) and lateral failures independently (P = .006). An increased risk for the presence of a rare allele for SNP rs17583842 was present in lateral failures compared with those that healed (P = .005). CONCLUSIONS Individuals with a family history of rotator cuff tearing were more likely to have repair failures. Significant association of a SNP variant in the ESRRB gene was also observed with lateral failure.

Significant evidence for linkage to chromosome 5q13 in a genome-wide scan for asthma in an extended pedigree resource

Asthma is a multifactorial disease with undetermined genetic factors. We performed a genome-wide scan to identify predisposition loci for asthma. The asthma phenotype consisted of physician-confirmed presence or absence of asthma symptoms. We analyzed 81 extended Utah pedigrees ranging from three to six generations, including 742 affected individuals, ranging from 2 to 40 per pedigree. We performed parametric multipoint linkage analyses with dominant and recessive models. Our analysis revealed genome-wide significant evidence of linkage to region 5q13 (log of the odds ratio (LOD)=3.8, recessive model), and suggestive evidence for linkage to region 6p21 (LOD=2.1, dominant model). Both the 5q13 and 6p21 regions indicated in these analyses have been previously identified as regions of interest in other genome-wide scans for asthma-related phenotypes. The evidence of linkage at the 5q13 region represents the first significant evidence for linkage on a genome-wide basis for this locus. Linked pedigrees localize the region to approximately between 92.3–105.5 Mb.