Forward/reverse reads support
This filter is recommended in all cases where an even distribution of forward and reverse reads at every position is expected. However, it should not be used for data sets such as large amplicons, where the ends of an amplicon are likely to be covered by only forward or reverse reads.
Due to sequencing or PCR artifacts and mapping issues, there can be some positions in the reference genome where only forward or only reverse reads are aligned. This can lead to certain alleles being present on one strand only.
If there is a strand bias from sequencing visible in the quality output check after sequencing, these should be regarded as suspicious regions that should be ignored during variant calling. If the user has selected the forward/reverse read support option, only variants that have a forward/reverse read balance of at least 0.05 are reported.
The forward/reverse balance is calculated as:
![$ \displaystyle Min(($](img199.gif)
![$ forward/$](img200.gif)
![$ total)($](img201.gif)
![$ reverse/$](img202.gif)
![$ total))$](img203.gif)
where
#forward = number of forward reads supporting the variant
#reverse = number of reverse reads supporting the variant
#total = all reads supporting the variant