Description Usage Arguments Details Value Functions See Also
Uses the method of moments and sets the ctrl parameters to
their previous defaults, i.e. include.mean.var=TRUE
,
obs.min.ingroup=2
and use.nonconv.groupest=TRUE
.
Estimates the parameters efficiency and lambda0 separately for each group of UMIs (i.e, for example separately for each gene, each exon, each cell, ...). To reduce the noise of the raw group-specific parameters, the raw group-specific parameters are shrunken towards global estimates for efficiency and lambda0 that are specified
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 | gwpcrpois.mom.groupwise(
formula,
data,
threshold = 1,
molecules = 1,
loss = expression(p0),
ctrl = list()
)
gwpcrpois.groupest(
formula,
data,
method = "mom",
threshold = 1,
molecules = 1,
loss = expression(p0),
ctrl = list()
)
|
formula |
Formula of the form |
data |
a |
threshold |
minimal number of observations a molecular family must have to count as unambiguously detected. Setting this to a value v >= 0 conditions the distribution on c >= c, i.e every value of c less than that gets assigned probability zero. |
molecules |
initial copy number |
loss |
an expression specifying how the loss, i.e. the percentage of all molecules (or UMIs) that was not observed, or removed by the read count threshold. In the simple case of each read count observation representing a separate molecules, the default value p0 is correct – the lost molecules are then simply those whose read count lies below the specified threshold. In more complex scenarios, e.g. if a single molecule produces separate read count for each strand, which are then either both rejected or both accepted, the additional rejection cases must be considered by a custom loss expression |
ctrl |
a list of settings controlling the estimation procedure.
Difference estimation methods recognize different possible ctrl
settings, unrecognized settings are ignored without warning. See
|
method |
the estimation method to use, either 'mle' for maximum likelihood estimation or 'mom' for method of moments. See Details. |
Apart from the ctrl parameters described in
gwpcrpois.est
, the following settings can be modified
number of CPU cores to use to compute group-specific estimates.
If set to a value greather than 1, the parallel
package must be
loaded.
how many observed per-UMI read counts a group must contain for group-specific estimates to be computed and used. Default is 5.
whether to use global estimates that didn't
fully converge (i.e. where gwpcrpois.est
reported a value other
than 0
for convergence). Default is FALSE
.
whether to use group-specific estimates
that didn't fully converge (i.e. where gwpcrpois.est
reported
a value other than 0
for convergence). Default is FALSE
.
whether to include columns with the group-wise
global mean and variance of readcounts in the output table. Default is
FALSE
.
whether to output progress and diagnostic messages, default
is FALSE
.
a data.table
containing one row per group, and
which besides the group key column(s) as indicated by the formula,
contains the following columns
efficiency |
the final group-specific estimate of the PCR efficiency |
lambda0 |
the final group-specific estimate of the number of reads per molecule |
loss |
the final group-specific estimate of the fraction of lost molecules. |
n.tot |
the final group-specific estimate of the total (loss-corrected) number of molecules. |
gwpcrpois.mom.groupwise
:
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.