CONAN: copy number variation analysis software for genome-wide association studies

Abstract

Background: Genome-wide association studies (GWAS) based on single nucleotide polymorphisms (SNPs) revolutionized our perception of the genetic regulation of complex traits and diseases. Copy number variations (CNVs) promise to shed additional light on the genetic basis of monogenic as well as complex diseases and phenotypes. Indeed, the number of detected associations between CNVs and certain phenotypes are constantly increasing. However, while several software packages support the determination of CNVs from SNP chip data, the downstream statistical inference of CNV-phenotype associations is still subject to complicated and inefficient in-house solutions, thus strongly limiting the performance of GWAS based on CNVs.

Results: CONAN is a freely available client-server software solution which provides an intuitive graphical user interface for categorizing, analyzing and associating CNVs with phenotypes. Moreover, CONAN assists the evaluation process by visualizing detected associations via Manhattan plots in order to enable a rapid identification of genome-wide significant CNV regions. Various file formats including the information on CNVs in population samples are supported as input data.

Conclusions: CONAN facilitates the performance of GWAS based on CNVs and the visual analysis of calculated results. CONAN provides a rapid, valid and straightforward software solution to identify genetic variation underlying the 'missing' heritability for complex traits that remains unexplained by recent GWAS. The freely available software can be downloaded at http://genepi-conan.i-med.ac.at.

Figure 1

Software Architecture. (A) In traditional approaches the application requests data, processes it locally and uploads the results. Thus, additional amount of data transfer reduce the performance. (B) The two-tier architecture of CONAN outsources all data intensive algorithms on the database server. The client invokes stored procedures to execute them on the database server itself; thus no upload of data is required and the client retrieves only the informative results.

Figure 2

CNVR Determination. (A) The boundaries of a sub-CNVR are determined using the start and end SNP of each CNV. (B) The basic algorithm designates a sub-CNVR as a CNVR if its frequency is greater than the threshold. (C) The extended algorithm merges consecutive sub-CNVRs and builds a single one on their basis.

Figure 3

Extended CNVR-Determination. The longest CNV of a subject in the CNVR with a percentage greater than the threshold (e.g. 50%) is used to define the final state.

Figure 4

Overview of steps. The CONAN analysis process is divided into three main steps: data import, data analysis, and data visualization.

Figure 5

Graphical User Interface. All imported datasets, their calculated CNV regions and associated analyses are organized in a tree structure. All CNV regions of the current selected dataset or association analysis are shown as a table: Genome-wide significant CNVRs are highlighted in red and regions with already know associations from SNP-GWAS http://www.genome.gov/gwastudies/ are highlighted in yellow. Information about a specified CNVR is listed in a separate dialog.

Figure 6

Data Visualization. CONAN supports the analysis process by several visualizations: (A) Visualization of the distribution of all detected CNV regions on each chromosome. (B) Visualization of associations via Manhattan plot enables a rapid identification of genome-wide significant CNVRs.