The coverage analysis tool is designed to identify regions in read mappings with unexpectedly low or
high coverage. Such regions may be indicative of a deletion or an amplification in the sample
relative to the reference. The algorithm fits a Poisson distribution to the observed coverage in the
positions of the mapping. This distribution is used as the basis for identifying the regions of
'Low coverage' or 'High coverage'. The user chooses two parameter values in the wizard: (1) a 'Minimum length' and
(2) a 'P-value threshold' value. The algorithm inspects the coverage in each of the positions in the read mapping
and marks the ones with coverage in the lower or upper tails of the estimated Poisson distribution, using the provided
p-value as cut-off. Regions with consecutive positions marked consistently as having low (respectively high) coverage,
longer than the user specified 'Minimum length' value are called as 'Low coverage' (respectively 'High coverage') regions.
The coverage analysis tool produces an annotation track carrying the name of the original file followed by (COV). This file can be visualized as an annotation track or as a table depending on the users choice. The annotation table contains a row for each detected low or high coverage region, with information describing the location, the type and the p-value of the detected region. The p-value of a region is defined as the average of the p-values calculated for each of the positions in the region.
The tool can also provide a report made of 2 tables. The first one, called References, lists per chromosome the number of reads, their length, and how many signatures of unexpectedly low or high coverage was found in the mapping. The second table lists on 2 rows low and high coverage signatures found, as well as how many reads were used to calculate these signatures.