What settings can be changed in Metaphlan2 to relax the parameters for classification? For example, what effect does changing --stat_q and --min_cu_len have on the profiling?
--stat_q is a parameter used by the robust average of the marker abundance. By default, this value in MetaPhlAn2 is set at 0.1, meaning that when the vector of abundances of each marker is sorted, only the marker abundances from the 10th to 90th percentile are used to compute the abundance of the clade.
By increasing this value MetaPhlAn will consider an higher number of markers, but it is possible that the number of False Positives and False Negatives increases.
min_cu_len is the minimum total nucleotide length of the markers in a clade for estimating the abundance without considering sub-clade: if the length of all the markers identified at the genus level is below 2000, and markers at the species level are available, the sum of the latter will be used to calculate the relative abundance at the genus level.