Could someone familiar with deML explain the 'rgqual' parameter a bit better? The default value listed for rgqual in the documentation is 80, but with the quality scoring system in all current Illumina data (phred 33), quality scores this high don't exist. Is this a different type of quality score? If so, could you elaborate a little on how it relates to PHRED quality, how it is calculated, and why the default is 80?
The rgqual refers to the Z_0 score. Z_0 is the likelihood for the top (most likely) read group on a PHRED scale. Considering the sequence of this read group as the template to sequence and using the principle of independence between bases, this likelihood is computed as:
The higher the probability of assignment, the lower this score becomes. If say the probability is almost 1 (perfect), the score Z_0 will be around 0. Conversely, if the probability of assignment is dismal around 0, the score will be 200 or 300. To compute this, we use quality scores but the correspondence is not 1 to 1, it depends on the template and sequenced bases.
I think the confusion stems from the use of the PHRED scale. PHRED is a mathematical formula and does not always refer to quality scores. Mapping quality is computed on a PHRED scale for instance.
