Limits...
Walking the interactome for candidate prioritization in exome sequencing studies of Mendelian diseases.

Smedley D, Köhler S, Czeschik JC, Amberger J, Bocchini C, Hamosh A, Veldboer J, Zemojtel T, Robinson PN - Bioinformatics (2014)

Bottom Line: Here, we analyze protein-protein association (PPA) networks to identify candidate genes in the vicinity of genes previously implicated in a disease.The analysis, using a random-walk with restart (RWR) method, is adapted to the setting of WES by developing a composite variant-gene relevance score based on the rarity, location and predicted pathogenicity of variants and the RWR evaluation of genes harboring the variants.Benchmarking using known disease variants from 88 disease-gene families reveals that the correct gene is ranked among the top 10 candidates in ≥50% of cases, a figure which we confirmed using a prospective study of disease genes identified in 2012 and PPA data produced before that date.

View Article: PubMed Central - PubMed

Affiliation: Mouse Informatics Group, The Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridgeshire CB10 1SA, UK, Institute for Medical Genetics and Human Genetics, Charité-Universitätsmedizin Berlin, Augustenburger Platz 1, 13353 Berlin, Genome Informatics Department, Institute of Human Genetics, University Hospital Essen, University of Duisburg-Essen, Hufelandstr. 55, 45122 Essen, Germany, McKusick-Nathans Institute of Genetic Medicine, John Hopkins University School of Medicine, Baltimore, MD 21205, USA, Department of Mathematics and Computer Science, Institute for Bioinformatics, Freie Universität Berlin, Takustrasse 9, 14195 Berlin, Germany, Institute of Bioorganic Chemistry, Polish Academy of Sciences, 61-701 Poznan, Poland, Berlin-Brandenburg Center for Regenerative Therapies, Charité-Universitätsmedizin Berlin, Augustenburger Platz 1, 13353 Berlin and Max Planck Institute for Molecular Genetics, Ihnestrasse 73, 14195 Berlin, Germany.

Show MeSH

Related in: MedlinePlus

Performance of ExomeWalker using STRING v9.05 as the source of interactome data. The bars show the percentage of exomes where the true disease gene is identified as the top hit or in the top 10 or 50 results. Either in-house or 1000 Genomes Project exomes were used. All exomes are filtered to remove synonymous, intergenic and intronic variants except for those in splice sites. In addition, variants with a MAF > 1% are excluded. Results are shown without (All) or with an AD or AR inheritance model applied. Ranking is either by Variant scoring that combines MAF and predicted pathogenicity, RWR analysis alone or ExomeWalker scoring that additionally includes evidence of protein–protein associations with other genes linked to the disease
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4221119&req=5

btu508-F1: Performance of ExomeWalker using STRING v9.05 as the source of interactome data. The bars show the percentage of exomes where the true disease gene is identified as the top hit or in the top 10 or 50 results. Either in-house or 1000 Genomes Project exomes were used. All exomes are filtered to remove synonymous, intergenic and intronic variants except for those in splice sites. In addition, variants with a MAF > 1% are excluded. Results are shown without (All) or with an AD or AR inheritance model applied. Ranking is either by Variant scoring that combines MAF and predicted pathogenicity, RWR analysis alone or ExomeWalker scoring that additionally includes evidence of protein–protein associations with other genes linked to the disease

Mentions: Figure 1 shows the results of our analysis using STRING v9.05 as the source of interactome data. Analysis was performed by adding known disease variants to either in-house exomes or 1000 Genome Project exomes and with and without the appropriate inheritance model for the disease being tested.Fig. 1.


Walking the interactome for candidate prioritization in exome sequencing studies of Mendelian diseases.

Smedley D, Köhler S, Czeschik JC, Amberger J, Bocchini C, Hamosh A, Veldboer J, Zemojtel T, Robinson PN - Bioinformatics (2014)

Performance of ExomeWalker using STRING v9.05 as the source of interactome data. The bars show the percentage of exomes where the true disease gene is identified as the top hit or in the top 10 or 50 results. Either in-house or 1000 Genomes Project exomes were used. All exomes are filtered to remove synonymous, intergenic and intronic variants except for those in splice sites. In addition, variants with a MAF > 1% are excluded. Results are shown without (All) or with an AD or AR inheritance model applied. Ranking is either by Variant scoring that combines MAF and predicted pathogenicity, RWR analysis alone or ExomeWalker scoring that additionally includes evidence of protein–protein associations with other genes linked to the disease
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4221119&req=5

btu508-F1: Performance of ExomeWalker using STRING v9.05 as the source of interactome data. The bars show the percentage of exomes where the true disease gene is identified as the top hit or in the top 10 or 50 results. Either in-house or 1000 Genomes Project exomes were used. All exomes are filtered to remove synonymous, intergenic and intronic variants except for those in splice sites. In addition, variants with a MAF > 1% are excluded. Results are shown without (All) or with an AD or AR inheritance model applied. Ranking is either by Variant scoring that combines MAF and predicted pathogenicity, RWR analysis alone or ExomeWalker scoring that additionally includes evidence of protein–protein associations with other genes linked to the disease
Mentions: Figure 1 shows the results of our analysis using STRING v9.05 as the source of interactome data. Analysis was performed by adding known disease variants to either in-house exomes or 1000 Genome Project exomes and with and without the appropriate inheritance model for the disease being tested.Fig. 1.

Bottom Line: Here, we analyze protein-protein association (PPA) networks to identify candidate genes in the vicinity of genes previously implicated in a disease.The analysis, using a random-walk with restart (RWR) method, is adapted to the setting of WES by developing a composite variant-gene relevance score based on the rarity, location and predicted pathogenicity of variants and the RWR evaluation of genes harboring the variants.Benchmarking using known disease variants from 88 disease-gene families reveals that the correct gene is ranked among the top 10 candidates in ≥50% of cases, a figure which we confirmed using a prospective study of disease genes identified in 2012 and PPA data produced before that date.

View Article: PubMed Central - PubMed

Affiliation: Mouse Informatics Group, The Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridgeshire CB10 1SA, UK, Institute for Medical Genetics and Human Genetics, Charité-Universitätsmedizin Berlin, Augustenburger Platz 1, 13353 Berlin, Genome Informatics Department, Institute of Human Genetics, University Hospital Essen, University of Duisburg-Essen, Hufelandstr. 55, 45122 Essen, Germany, McKusick-Nathans Institute of Genetic Medicine, John Hopkins University School of Medicine, Baltimore, MD 21205, USA, Department of Mathematics and Computer Science, Institute for Bioinformatics, Freie Universität Berlin, Takustrasse 9, 14195 Berlin, Germany, Institute of Bioorganic Chemistry, Polish Academy of Sciences, 61-701 Poznan, Poland, Berlin-Brandenburg Center for Regenerative Therapies, Charité-Universitätsmedizin Berlin, Augustenburger Platz 1, 13353 Berlin and Max Planck Institute for Molecular Genetics, Ihnestrasse 73, 14195 Berlin, Germany.

Show MeSH
Related in: MedlinePlus