NextGENe Online Help : NextGENe Viewer Comparison Reports and Tools : The CNV (Copy Number Variation) Tool : SNP-based Normalization with smoothing
SNP-based Normalization with smoothing
TheSNP-based Normalization with Smoothing coverage option has three components – the Log2 ratio calculated based on the perfect heterozygote SNP positions, the score, and the Log2 ratio based on the SNP positions for adjacent (neighbor) regions.
• Log2 ratio calculated based on the perfect heterozygote SNP positions—The CNV tool checks the coverage for at least three positions in each region. Perfect heterozygote SNP positions, which are positions with a user-specified variant frequency in the selected regions in at least one sample, are chosen first. If three perfect heterozygote SNP positions are not found, the tool chooses positions every 100 bp, starting in the middle of the region. If there are more than 100 bp without a Perfect heterozygote SNP position, the tool chooses additional positions every 100 bp. The tool then calculates the median coverages for these positions and normalizes the median coverage values relative to the global coverage. The Log2 ratio of the normalized coverage values of the two samples is then calculated.
• Score – A Phred-scaled score is calculated for each potential call (duplication, deletion, and normal) based on a binomial distribution that considers the coverage.
• Log2 ratio for adjacent (neighbor) regions – Considers the Log2 ratio calculated based on SNP positions for the three regions that are directly upstream and the three regions that are directly downstream of the current region.
CNV calls are made according to the following:
  
Component Values
Call
Upstream and downstream neighbor log2 ratio and current log2 ratio = 0
Uncalled
Log2ratio > 20
Duplication
Log2ratio < -20
Deletion
Log2ratio > 2 and duplication score > 20
Duplication
Log2ratio < -2 and deletion score > 20
Deletion
Upstream and downstream neighbor log2 ratios > 0.4 and duplication score > 10
Duplication
Upstream and downstream neighbor log2 ratios < -0.5 and deletion score > 10
Deletion
Neighbor called as a duplication and upstream, downstream and current log2 ratios > 0.3
Duplication
Neighbor called as a Deletion and upstream, downstream and current log2 ratios < -0.4
Deletion
Upstream, downstream, and current log2 ratios are > -0.5 and < 0.4
Normal
The median of upstream, downstream and current log2 ratios > 0.4 and duplication score > 10
Duplication
The median of upstream, downstream and current log2 ratios < -0.5 and deletion score > 10
Deletion
The median of upstream, downstream and current log2 ratios < 0.4 and > -0.5
Normal
Neighbor called as a Duplication and duplication score > 1
Duplication
Neighbor called as a Deletion and deletion score > 1
Deletion
Neighbor called as Normal and normal score > deletion score and > duplication score
Normal
If none of the above criteria are met, then Uncalled, unless:
 
• If Uncalled and the coverage for the sample and the control > 1000x, the current log2 ratio > 0.5, and the duplication score > 100
• Duplication
• If Uncalled and the coverage for the sample and the control > 1000x, the current log2 ratio < -0.9 and the deletion score > 100
• Deletion