SNiPA : an interactive, genetic variant-centered annotation browser

1Genome-wide association studies (GWAS) and next-generation sequencing (NGS) are performed routinely to identify genetic variants and novel genes implicated in both common and rare human diseases. A key step in translating results from such studies into a better understanding of molecular disease mechanisms and, ultimately, into clinical applications, is the prioritization of potentially functional variants that may be active in vivo. To this end, comprehensive collection and evaluation of existing functional annotation from genetic, informatics and experimental resources is essential. This comprises the integration of data and knowledge across multiple levels including the variant, the gene and the chromatin level.

2Several large resources (Ensembl, UCSC, NCBI, etc.) aim at providing genome-wide genome-level annotation tracks from an extensive set of resources. However, retrieving statistical and functional annotation relevant at the single nucleotide level remains difficult. For instance, common genome browsers often display single nucleotide variants (SNVs) as thin bars that trail away in the wealth of other annotation tracks and are even less prepared to display statistics such as linkage disequilibrium (LD) relationships between variants. This limits visual distinction of relevant variants from those without relevant annotations and leaves the complex task of aggregating position-based data to the researcher. Variant-centered resources, on the other hand, typically concentrate on specific types of data such as amino acid changes, expression quantitative trait loci (eQTLs), trait associations or regulatory effect predictions. Moreover, these annotations are often presented in resource-specific data structures.

3For individual inspection of single variants, both resource types are extremely valuable. However, for simultaneous processing of larger variant sets, collection and examination of annotations from different data sources quickly becomes cumbersome. This presents a major bottleneck in genome-wide scans of genetic influences on human traits since the collection of such evidences is the key to understanding the effects of phenotype-linked genetic variants.

4Here we propose SNiPA, a web service offering variant-centered genome browsing and interactive visualization tools tailored for easy inspection of many variants in their locus context (Fig. 1).

5SNiPA includes a wide range of genome-level datasets contained in the Ensembl database as an established backbone of annotations for the human genome. We combine this backbone with numerous variant-specific annotations taken from published datasets. Thus, SNiPA covers information ranging from regulatory elements, over gene annotations to variant annotations and associations (Table 1; Supplementary Text S1). SNiPA contains annotations for all bi-allelic variants in phase 3 version 5 of the 1000 genomes project and provides pre-calculated LD-data for r2 ≥ 0.1 for all super populations (African, American, South and East Asian, European). We use the Ensembl VEP tool for primary effect prediction of SNVs. Additional position-based data is included in the VEP prediction as custom annotation files. For other annotations, we wrote a Perl module to extend the output provided by VEP (Table 1; Supplementary Text S1).

6SNiPA provides user-friendly starting points for annotating individual SNVs as well as sets of SNVs, LD blocks or genetic regions of interest. We have implemented several entry points to access the data: (i) a variant-centered implementation of a genome browser (‘Variant Browser’); (ii) ‘Association Maps’ for browsing through GWAS results; (iii) an interface for batch retrieval of variant annotations via ID-list, gene ID or genomic coordinates (‘Variant Annotation’); (iv) a combined listing of annotations across a set of variants within LD blocks or chromosomal regions (‘Block Annotation’); (v) ‘Regional Association Plot’ and ‘Linkage Disequilibrium Plot’ that combine publication-ready plotting of association results and LD values, respectively, with the interactive interface of the ‘Variant Browser’; (vi) ‘Proxy Search’ and ‘Pairwise LD’ that allow querying precalculated LD values augmented with variant annotations. SNiPA enables the user to download condensed annotation data in tabular format for further off-line processing. Detailed descriptions of SNiPA modules are available in the online documentation and Supplementary Text S1.

7The complex information contained in SNiPA is organized in a clear, comprehensive and informative structure extending effect categories contained in the Sequence Ontology (Supplementary Text S1). For instance, variant annotations are presented as ‘SNiPAcards’ grouping information into semantic sections. All annotations are linked to their primary sources and to the Ensembl genome browser.

8Mechanistic characterization of variants identified by genetic studies is the key to understanding molecular disease mechanisms. SNiPA combines a comprehensive set of genomic annotations with a genetic variant-based genome browser to simplify the task of variant annotation. SNiPA as well as all underlying data is freely available to the scientific community (commercial use may be limited by third-party constraints) and will be automatically updated following the Ensembl releases.

9This work was supported by the Helmholtz Portfolio theme ‘Metabolic Dysfunction and common disease’ and by the research project Greifswald Approach to Individualized Medicine (GANI_MED) (BMBF: 03IS2061A). K.S. is supported by Biomedical Research Program funds at Weill Cornell Medical College in Qatar, a program funded by the Qatar Foundation.

10Conflict of Interest: none declared.