GenerateNoisyTumors
to only write signatures that
are present in exposures to CSV file.cosmicsig
and mSigAct
.
Due to fix in function GenerateSynExposureOneSample
to round the mutations
due to each signature in version 1.0.9, the wrapper functions to generate
data sets for Nature paper and SBS1-SBS5 paper fail to reproduce the data sets
used by these papers.
In SynSigGen 1.1.0, we disabled the rounding of mutations due to each signature when running generator functions for legacy data sets, so that these legacy data sets can be reproduced.
sig.params
in function GenerateSyntheticExposures
.URL
and BugReports
in DESCRIPTION.
ID.MMR.params
because the list of MSI-H tumors is not
complete when generating the old data.
Updated exported function MergeExposures
to sort the signature ids of the
merged exposure.
Suppressed warning messages that came from internal function NumFromId
.
GetMutationType
to check all the signature
names to determine the mutation type.
GetSynSigParamsFromExposures
and internal function
SynSigParamsOneSignature
to only use the size
parameter from empirical
signature parameters if there is only one data to fit the negative binomial
distribution. The mu
parameter will be using the original mutation count.
GenerateSyntheticTumors
to show informative messages
about regenerating parameters from the synthetic exposures and compare with that
from real exposures if verbose > 0
.Added new argument tumor.marker.name
in function GenerateSyntheticTumors
.
Added new argument sig.params
in exported functions
GetSynSigParamsFromExposures
and GenerateSyntheticTumors
.
Added two new exported functions GenerateListOfSigParams
and
GenerateSyntheticTumorsFromSigParams
.
Updated function GenerateSynExposureOneSample
to round the synthetic exposure to make
it biologically reasonable.
Updated function GetSynSigParamsFromExposures
not to drop rare signatures
that are present in only one sample when distribution
is neg.binom
. Instead,
this function will use the empirical signature parameters from all cancer types.
Updated function GenerateSyntheticTumors
to enable generating different number
of synthetic tumors for each cancer type.
Updated function GetSynSigParamsFromExposures
to drop signatures which do not
have empirical signature parameters from all cancer types.
Updated function GenerateSyntheticTumors
and GenerateNoisyTumors
to append
mutation type information to the names of files generated.
Updated function GenerateSyntheticExposures
to remove signatures that have zero
exposure in the synthetic data.
Added default value 100 for argument n.binom.size
in function AddNoise
.
Added new exported data signature.params
.
Added new exported data ID.MMR.params
.
Added two new internal functions NumFromId
and SortSigId
.
GenerateSynExposureOneSample
to round the mutations
due to each signature.
Added new exported function GenerateNoisyTumors
.
Minor change to function GenerateSyntheticTumors
for always writing the
signature names in the parameters CSV file in case some signature is not present
in the synthetic data.
Added new argument distribution
in functions SynSigParamsOneSignature
and
GetSynSigParamsFromExposures
to enable fitting the exposures using negative
binomial distribution.
Added new argument distribution
and sig.params
in function
GenerateSynExposureOneSample
to enable generating synthetic exposures using
negative binomial distribution.
Added new argument distribution
in function GenerateSyntheticExposures
to
enable generating synthetic exposures using negative binomial distribution.
Added new exported function GenerateSyntheticTumors
which enables generating
synthetic tumors using negative binomial distribution.
Added new argument verbose
in function GenerateSyntheticTumors
which can
cat various messages if value greater than 0.
Added new argument cancer.type
in function GetSynSigParamsFromExposures
.
SynSigParamsOneSignature
to generate parameter
estimates (negative binomial distribution) for one signature only using those
tumors which have mutations for this particular signature.SynSigParamsOneSignature
to return NA for the parameter
estimates (negative binomial distribution) for one signature if there is only
one tumor which has mutation for this particular signature.
Added data object "SBS1SBS5datasetNames" as names of SBS1-SBS5 spectra datasets generated by CreateSBS1SBS5CorrelatedSyntheticData()
Added new argument sig.matrix
in functions GenerateSynExposureOneSample
and GenerateSyntheticExposures
so that we can resample the synthetic exposures to make
sure the total mutations for the reconstructed catalog is not zero.
GetSynSigParamsFromExposures
when the exposures only have one row that are
not all zeros.CreateSBS1SBS5CorrelatedSyntheticData()
-
The files of synthetic datasets are under the dataset folders, and no more
sp.sp
folders.S.0.1.Rsq.0.1
)sp.sp
, means that two ground-truth signatures (SBS1
and SBS5
) are from results of SigProfiler in (https://doi.org/10.1038/s41586-020-1943-3)README
Create.3.4.40.Abstract()
to Create.3.5.40.Abstract()
R/SigSimilarity.R
(functionality now in ICAMSxtra
)
and adjusted code in this package.AddNoise
.biocViews
as a field name in package DESCRIPTION.Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.