Background
Recent advances in microarray and sequencing technology have enabled the identification of genomic variation in humans. Among the different types of variants, Single Nucleotide Polymorphisms (SNPs) and Copy Number Variation (CNV) have attracted attention due to the relatively frequency of their incidence in the human genome
The study of small CNVs may have been limited by the shortcomings of previously used platforms such as SNP genotyping and Comparative Genomic Hybridization array (aCGH)
Comparison of Genovar with previously reported CNV analysis tools
Plat-
form
Module
CGHPRO
VAMP
ISACGH
MD-SeeGH
SIGMA2
CGcgh
CGHweb
CHESS
SnoopCGH
CNA-Reporter
SEURAT
Wavi-CGH
FISH Oracle
CNVAS
GENOVAR
Year
2005
2006
2007
2008
2008
2008
2008
2009
2009
2010
2010
2010
2011
2011
2011
CNV detection
O
O
O
O
O
O
O
O
O
O
O
O
O
CGH
CNV visualization
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
CNV quality control
O
O
Annotation
O
O
O
O
O
O
O
O
O
O
O
O
O
O
CNV detection
O
NGS
CNV visualization
O
CNV quality control
O
Annotation
The present study describes the development of a JAVA-based stand-alone program called Genovar that efficiently detects CNV regions and provides a visual inspection function to reduce false positive CNV calls based on aCGH and NGS data. Genovar can analyze both aCGH and NGS data (Table
Genovar consists of three major components. Firstly, Genovar visualizes aCGH data and sequence alignment of chromosomal regions. Regarding NGS data, Genovar provides a read-depth plot, and summary information of each read when a certain read is selected in the panel. With respect to CNVs, the SW-array algorithm
Results and discussion
System architecture and features
Genovar is a stand-alone application for the identification and visualization of CNV regions based on two major types of data input such as aCGH and Binary Alignment/Map (BAM) file formats. The graphical user interface was implemented by JAVA swing, and user interactions are handled by this intuitive interface. The Smith-Waterman Array (SW-ARRAY) algorithm
System architecture of Genovar
System architecture of Genovar. Genovar mainly takes comparative genomic hybridization array (aCGH) and binary sequence alignment/map (BAM) file as mandatory inputs for copy number variation (CNV) and next generation sequencing (NGS) analysis. Optional inputs include probe-gene mapping information and reference sequence (UCSC fastA format). Using two major modules (aCGH and NGS), Genovar performs analyzing tasks such as known variation reporting, graphical visual inspecting and variation detecting.
Genovar user's guide. This file is the user's guide of Genovar, which was uploaded to the Genovar website
Click here for file
• Analyzing/comparing them with multiple aCGH samples and windows.
• Elimination of spurious signals, and statistical block operation of log2 ratio values.
• Identification of CNV regions using the SW-Array algorithm and thresholds.
• Notation of known CNVs in comparison to newly identified CNV regions using DGV and dbSNP.
• Graphic representation of CNV region and aCGH information with multiple samples.
The features of the NGS module are as follows.
• User-intuitive and fast navigation on chromosomes for retrieving reads in terms of locus range query.
• Graphic-based display of sequence alignment results and read information in BAM (binary sequence alignment/map) format.
• Calculation of read-depth and allele frequency of each locus in the alignment area specified by the user and identification of known SNPs and CNV from dbSNP and DGV, respectively.
• Comparison of sequence alignment results between different samples.
System inputs
As mentioned earlier, Genovar uses two major input data formats such as aCGH or NGS. First, the aCGH input format includes probe ID and name, chromosome number, probe starting and ending locus, and a series of log2 intensity values delimited by 'TAB'. The aCGH sample formats are available on the Genovar website
Chromosomal view from array CGH data
After loading an aCGH input file, Genovar displays a view of the aCGH value corresponding to the first sample in the file in a whole-chromosome context (Figure
Chromosomal view and statistical summary
Chromosomal view and statistical summary. A. Whole-chromosome scale view of the first sample in the file. Duplication and deletion regions are represented by green and red color, respectively. B. A series of log ratio values for a specific chromosome are given in table form, and a plot of each log ratio value is also offered, along with a cytoband. Zoom in and out functions are also included, via a pop-up menu. C. Statistical summary including mean, median, max, min, sum, and variance, as well as an outlier filter, for a particular region, in single-chromosome view.
Elimination of spurious signals
As mentioned before, due to limitations such as low resolution of array platforms, platform specificity, and the type of CNV class, current CNV analysis tools often produce false positive results. For the CNV detection process, the segmental mean of log2 value of probes in the CNV region is the most important value for defining the CNV region. Moreover, noisy log2 values of probes in a CNV region may hinder the discrimination of a spurious signal from a true signal. Therefore, it is necessary to manually separate spurious signals from true signals by visual inspection.
To this purpose, Genovar provides a statistical summary pertaining to the region of interest (Figure
Copy number detection and reporting known CNV region
Genovar detects copy number variant regions using the Smith-Waterman Array (SW-ARRAY) algorithm
Copy number variation (CNV) detection
Copy number variation (CNV) detection.
Genomic comparison with multiple array CGH samples
Another useful function of Genovar is the comparison of CNV regions or aCGH values between samples. Genovar displays CNV regions in a specific chromosomal view named heat map (Figure
Comparison with multiple samples
Comparison with multiple samples. A. Heat map of copy number variation (CNV) regions and detailed view of specific CNV regions. This view is obtained by setting the starting and ending position using a pop-up menu. A heat map is shown for each chromosome. B. Comparative genomic hybridization (CGH) values and comparison of samples. Scatter dots represent the log2-scale signal intensity ratio between sample and reference. The colored samples correspond directly to the same colored spot on the cytoband view.
Displaying sequence alignments and read information from a BAM file
The recent NGS technology allows the discovery of small CNVs. In particular, read-depth coverage of NGS data is a very useful resource to detect homogeneous deletions. Genovar imports BAM formats and displays sequence alignment results. The alignment view (Figure
Sequence alignment view of a single binary sequence alignment/map (BAM) file
Sequence alignment view of a single binary sequence alignment/map (BAM) file. A single alignment view contains cytoband information, locus range, read-depth coverage of each locus, zoom level, and frequency of nucleotides for the current viewing window.
Genetic variant inspection based on sequence alignment results
Genovar shows multiple sequence alignment results simultaneously (Figure
Sequence alignment comparison of two binary sequence alignment/map (BAM) files
Sequence alignment comparison of two binary sequence alignment/map (BAM) files. Comparison of sequence alignment results from one sample with those from another sample. Results from dbSNP in a user-specified locus range are a useful means of detecting unknown SNPs.
Conclusions
Genovar is a useful tool to detect the CNV with the robust SW-algorithm as a dynamic programming solution and to visualize the detailed CNV information for performing quality control process. The user-friendly graphical interface of Genovar enables the user to identify CNV regions more efficiently. This graphical user interface was implemented by JAVA swing. Moreover, Genovar utilizes two major types of data formats such as those of aCGH or NGS to visualize CNV regions. Genovar compares the detected CNV regions with previously reported CNV regions that have been deposited in DGV/dbSNP. These functions are especially powerful because they allow users to verify whether the CNV regions and SNPs found in their own dataset are truly novel, dramatically reducing time and effort.
Genovar has two distinct advantages over previously reported softwares. It enables users to eliminate spurious ones from true signals through visual inspection and summarized information of detected CNV. Moreover, it even visualizes sequence alignment along with chromosomal regions. For sequence alignment data, Genovar provides a read-depth plot and summarized information of each read when a certain read is selected in the panel.
The visual inspection function of Genovar was used in many practical CNV analysis projects
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
KSJ was mainly responsible for the development and set-up of the software, performed the simulations, drafted the manuscript, and wrote most of the methods section. SM and YJK participated in writing the manuscript and designing and testing of the software, and contributed ideas regarding key operations of Genovar; BJK participated in designing and testing of the software. KP revised the manuscript and coordinated the project. All authors read and approved the final manuscript.