In cancer research, the widely accepted multiple-hit hypothesis states that cancer is the result of accumulated mutations, and is dependent on both the activation of proto-oncogenes and deactivation of tumor suppressor genes. Thus, after identifying a change of interest in your data set, whether it is a copy number change or a sequence variant, one of the next questions is often: what other changes are associated with this aberration? To answer this question, Nexus Copy Number 7 has a tool, Concordance function (found in the data set tab under the Tools button), which identifies statistically significant copy number alterations and sequence variants which co-occur with the change of interest within your data set.
To use the concordance function, you first need to specify the initial aberration of interest. This can be either a gene or a region, and can include any abnormality described in Nexus: Gain, Loss, LOH (loss of heterozygosity) or sequence variant. You will also indicate whether it is okay for the specified aberration to occur anywhere within the region or only to be fully covered by selecting within a check box. Note that only one gene/region can be queried at a time. Nexus will then query the samples selected within your data set to see if that specific aberration has been detected. As long as at least one sample has the genomic change queried, a Fisher Exact Test is then performed to compare the two groups (those with the specified aberration and those without) and all statistically significant regions of change are listed.
As an example, let’s say I have noticed a high percentage of BRAF mutations within my data set. I can use the concordance function to identify statistically significant changes that co-occur with a BRAF mutation. I can select for any mutation within the gene, by selecting gene and typing in BRAF, or I can focus on mutations solely within the BRAF V600 activating hotspot locus, by selecting region and typing in chr7:140453135-140453137. I would not select Completely Covered, as this would require samples to have sequence variants along the entire span of the gene or region. I can also include a custom label for this query.
Upon clicking OK, the software is querying every sample selected from my initial data set and dividing them into two groups: those which have a sequence variant anywhere within the region I specified versus those that do not have a sequence variant within the region. A Fisher’s exact test is then performed between these two groups and a report is generated with the regions of statistically significant alterations (copy number, allelic and sequence variant), with options for adjusting the p-value threshold and the differential threshold. This table can also be tailored to display additional information by selecting Modify View; in this example I have included a list of genes, miRNA and Sanger census genescontained within the region.
I now have the option of displaying these results in a circular plot. I have selected the co-occurring changes of interest in the right hand box, which include Sanger census genes, miRNAs and highly differential regions between the two groups. The Label column allows me to select what is displayed; in this example I have selected the gene name.
Thick lines indicate concordance (frequent among the sample group with BRAF V600 mutation) while dashed-lines indicate anti-concordance (frequent among the samples WT at BRAF V600). Red indicates loss, blue indicates gain and black indicates sequence variant. The track line indicates the p-value for each region selected. In the example shown below, BRAF V600 mutation was queried (Q) and several significant regions (A-M) are displayed.
I also have the option of visualizing the comparison testing between the two groups in a linear chromosome view, by selecting the Genome Tab.
This tool is just one way we can identify significant alterations of interest among sequence and copy number data.