See
also
Octave plot
Log-normal abundance distribution
otutab_octave command
Distribution shape
If the shape of
the distribution looks like a bell curve or truncated bell curve, then the
distribution is approximately log-normal,
suggesting that the OTUs are participating in an interacting ecosystem.
Other mathematical distributions can be recognized if the shape is different
(Edgar and Flyvbjerg, manuscript in preparation).
Setting an OTU abundance cutoff
Octave plots can indicate a good choice of
OTU abundance cutoff.
Estimating
read depth (sampling effort)
Octave plots can indicate the
number of additional reads needed to
capture missing OTUs.
Incomplete sampling
If the
distribution appears to be truncated, e.g. if part of the left side of the
bell curve is missing, his indicates that the full diversity has not been
sampled. See log-normal distribution for an
example.
Extrapolating richness
If a clear
peak is visible, then on the assumption that the distribution is
approximately symmetrical, this suggests that more than half of the
diversity has been sampled and it is reasonable to extrapolate richness
using an estimator such as Chao1, mirror or FE. If the peak is close to the
y axis or absent, then it is probably not justified to
extrapolate because there is not enough data.