Signal intensity data: LRR (Log R Ratio) and BAF (B Allele Frequency) of each individual and each probe

Additional input files for PennCNV as described in its manual: PFB (Population Frequency of B allele), HMM, and GCModel files

Linux environment with PennCNV installed: we assume the user has PennCNV installed or has the knowledge on how to obtain and install the software; more information is available on the PennCNV Web site (http://www.openbioinformatics.org/penncnv/penncnv_installation.html)

GenomeStudio or BeadStudio (Illumina) for exporting signal intensity files from Illumina SNP array project files

We encourage the reader to browse the respective software package Web sites (provided at the end of this unit) to find out more details on hardware requirements. In general, modern PCs and Linux servers with 2 to 4 GB RAM should be sufficient for running the programs we use in this unit. Analysis of larger datasets (on the scale of thousands subjects) may require more storage space and memory.

Illumina recommends running GenomeStudio (formerly BeadStudio) on a Windows (XP or later) computer with Intel Celeron Duo or faster 64‐bit CPU, at least 8 GB memory, at least 100 GB storage space, and 1280 × 1024 screen resolution for better viewing

PennCNV runs on Linux systems. Both source code and pre‐compiled executables are available. Instructions for installation on Windows systems with Cygwin or ActivePerl are provided on the PennCNV Web site. See Time Considerations for processing large datasets.

Precompiled executables for Windows and Linux systems are available for both R and PLINK

Output file from the PennCNV software (see protocol 1 ) that contains all called CNVs

Individuals' case/control status and phenotypes or factors that may confound the relation between CNVs and the disease state; these pieces of information are equivalent to those of PLINK FAM and covariate files

Linux environment with R installed. We assume the user has installed R or has the knowledge on how to obtain and install the software from the R Web site (http://www.r‐project.org/). Comprehensive documentation is available there.

R script (penncnv2cnpr.r, which can be downloaded at http://www.currentprotocols.com/protocol/hg0127)

For hardware requirements, see protocol 1 materials list

Output file from the PennCNV software (see protocol 1 ) that contains all called CNVs

PLINK FAM file from the Genome‐Wide Association Study (GWAS) SNP data.

Optional files describing user‐specified genomic regions for burden tests. For example, a file containing the coordinates of all known genes on the human genome. Each row specifies one genomic region (chromosome, start, and end positions).

Linux environment with PLINK installed. We assume the user has installed PLINK or has the knowledge on how to obtain and install it (http://pngu.mgh.harvard.edu/∼purcell/plink/). Comprehensive documentation is available there.

For hardware requirements, see protocol 1 materials list

Output file from the PennCNV software (see protocol 1 ) that contains all called CNVs

A Web browser that is compatible with the UCSC Genome Browser.

Format CNV files into the BED format. A BED file is a tab‐delimited file that represents genomic features, such as genes or CNVs as integer intervals one interval per line, and describes how these intervals to be displayed on the UCSC browser as a custom track (see unit 18.6 ). Only the first three fields—chromosome/scaffold name, start position and end position—describing the genomic location are required but the optional fields, such as name, strand, etc., and the “track line,” make the visualization more informative. Please refer to http://genome.ucsc.edu/FAQ/FAQformat.html#format1 for more details. For this protocol, we put seven fields and a track line in one BED file. An example may look like this:track name=test1 description=test1 visibility=3 colorByStrand="255,0,0 0,0,255" useScore=0 Although the track line appears as two lines, it is in fact one single line. Both of the last two fields being identical to the second one, i.e., the start position, makes the bars representing the CNVs thinner so as to accommodate more CNVs in one given space. For a small number of items, a generic text editor or Excel is sufficient to do the conversion manually. For a large number, however, a program is usually needed to do it efficiently and correctly. The following is an example Perl script that reads a PennCNV output file and converts it into a BED file. The script gives the contrast between deletions and duplications by assigning a strand status to each CNV (‘+’ when CN < 2 and ‘−’ when CN > 2), and using the “colorByStrand” attribute in the track line. The two colors for the two strands are specified by RGB color codes and divided by a space. To visualize the contrast between cases and controls, then the coding for strand should be used to encode disease status instead, and thus duplications and deletions should be separated into two tracks:

• #!/bin/perl
• use strict;
• ## This script prints the output to STDOUT. Use redirect to output the results to a file.
• # check if track name and input filename are provided
• die "Usage: $0 trackname infile\n" if scalar @ARGV < 2;
• my ($track, $infile) = @ARGV;
• # print the track line
• printf("track name=$track description=$track visibility=3 colorByStrand=\"255,0,0 0,0,255\" useScore=0\n");
• # open the input file and start processing line by line
• open(FIN, $infile) ∥ die "cannot open $infile\n";
• while (<FIN>) {
• # split one line into fields (the delimiter can be one or multiple spaces)
• my @arr=split(/\s+/,$_);
• # further split the first field into chr and positions
• my @ele=split(/[:‐]/,$arr[0]);
• # convert to 0‐based position
• my $start = $ele[1] ‐ 1;
• # split the copy number field
• my @cn=split(/[,=]/,$ele[3]);
• # assign deletion (CN<2) to positive strand '+' and duplication '‐' printf("%s\t%d\t%d\t%s\t%s\t%d\t%d\n",$ele[0],$start,$ele[2],$arr[4],$cn[2] < 2 ? '+' : '‐', $start, $start)
• }
• close FIN;

For hardware requirements, see protocol 1 materials list