Usage¶
General usage¶
Here is a description of all possible parameters. The tool take around 1 minute for 1d, 2000 cells, default parameters, independently of the number of gaussian and around 30 seconds for 300 cells. For the 2d 300 cells is around 3 minutes, 2000 cells around 15 minutes. Increasing the number of samples in the MCMC or in the burning phase will increase the time.
baredSC_1d¶
Run mcmc to get the pdf for a given gene using a normal distributions. The full documentation is available at https://baredsc.readthedocs.io
usage: baredSC_1d [-h] (--input INPUT | --inputAnnData INPUTANNDATA)
--geneColName GENECOLNAME
[--metadata1ColName METADATA1COLNAME]
[--metadata1Values METADATA1VALUES]
[--metadata2ColName METADATA2COLNAME]
[--metadata2Values METADATA2VALUES]
[--metadata3ColName METADATA3COLNAME]
[--metadata3Values METADATA3VALUES] [--xmin XMIN]
[--xmax XMAX] [--xscale {Seurat,log}]
[--targetSum TARGETSUM] [--nx NX] [--osampx OSAMPX]
[--osampxpdf OSAMPXPDF] [--minScale MINSCALE]
[--nnorm NNORM] [--nsampMCMC NSAMPMCMC]
[--nsampBurnMCMC NSAMPBURNMCMC]
[--nsplitBurnMCMC NSPLITBURNMCMC] [--T0BurnMCMC T0BURNMCMC]
[--seed SEED] [--minNeff MINNEFF] [--force] --output OUTPUT
[--figure FIGURE] [--title TITLE]
[--removeFirstSamples REMOVEFIRSTSAMPLES]
[--nsampInPlot NSAMPINPLOT] [--prettyBins PRETTYBINS]
[--logevidence LOGEVIDENCE] [--coviscale COVISCALE]
[--nis NIS] [--version]
Named Arguments¶
- --version
show program’s version number and exit
Required arguments¶
- --input
Input table (tabular separated with header) with one line per cell columns with raw counts and one column nCount_RNA with total number of UMI per cell optionally other meta data to filter.
- --inputAnnData
Input annData (for example from Scanpy).
- --geneColName
Name of the column with gene counts.
- --output
Ouput file basename (will be npz) with results of mcmc.
Optional arguments to select input data¶
- --metadata1ColName
Name of the column with metadata1 to filter.
- --metadata1Values
Comma separated values for metadata1 of cells to keep.
- --metadata2ColName
Name of the column with metadata2 to filter.
- --metadata2Values
Comma separated values for metadata2 of cells to keep.
- --metadata3ColName
Name of the column with metadata3 to filter.
- --metadata3Values
Comma separated values for metadata3 of cells to keep.
Optional arguments to run MCMC¶
- --xmin
Minimum value to consider in x axis.
Default: 0
- --xmax
Maximum value to consider in x axis.
Default: 2.5
- --xscale
Possible choices: Seurat, log
scale for the x-axis: Seurat (log(1+targetSum*X)) or log (log(X))
Default: “Seurat”
- --targetSum
factor when Seurat scale is used: (log(1+targetSum*X)) (default is 10^4, use 0 for the median of nRNA_Counts)
Default: 10000
- --nx
Number of values in x to check how your evaluated pdf is compatible with the model.
Default: 100
- --osampx
Oversampling factor of x values when evaluating pdf of Poisson distribution.
Default: 10
- --osampxpdf
Oversampling factor of x values when evaluating pdf at each step of the MCMC.
Default: 5
- --minScale
Minimal value of the scale of gaussians (Default is 0.1 but cannot be smaller than max of twice the bin size of pdf evaluation and half the bin size).
Default: 0.1
- --nnorm
Number of gaussian to fit.
Default: 2
- --nsampMCMC
Number of samplings (iteractions) of mcmc.
Default: 100000
- --nsampBurnMCMC
Number of samplings (iteractions) in the burning phase of mcmc (Default is nsampMCMC / 4).
- --nsplitBurnMCMC
Number of steps in the burning phase of mcmc.
Default: 10
- --T0BurnMCMC
Initial temperature in the burning phase of mcmc (>1).
Default: 100.0
- --seed
Change seed for another output.
Default: 1
- --minNeff
Will redo the MCMC with 10 times more samples until the number of effective samples that this value (Default is not set so will not rerun MCMC).
- --force
Force to redo the mcmc even if output exists.
Optional arguments to get plots and text outputs¶
- --figure
Ouput figure filename.
- --title
Title in figures.
- --removeFirstSamples
Number of samples to ignore before making the plots (default is nsampMCMC / 4).
- --nsampInPlot
Approximate number of samples to use in plots.
Default: 100000
- --prettyBins
Number of bins to use in plots (Default is nx).
Optional arguments to get logevidence¶
- --logevidence
Ouput file to put logevidence value.
- --coviscale
Scale factor to apply to covariance of parameters to get random parameters in logevidence evaluation.
Default: 1
- --nis
Size of sampling of random parameters in logevidence evaluation.
Default: 1000
combineMultipleModels_1d¶
Combine mcmc results from multiple models to get a mixture using logevidence to infer weights.
usage: combineMultipleModels_1d [-h]
(--input INPUT | --inputAnnData INPUTANNDATA)
--geneColName GENECOLNAME
[--metadata1ColName METADATA1COLNAME]
[--metadata1Values METADATA1VALUES]
[--metadata2ColName METADATA2COLNAME]
[--metadata2Values METADATA2VALUES]
[--metadata3ColName METADATA3COLNAME]
[--metadata3Values METADATA3VALUES] --outputs
OUTPUTS [OUTPUTS ...] [--xmin XMIN]
[--xmax XMAX] [--xscale {Seurat,log}]
[--targetSum TARGETSUM] [--nx NX]
[--osampx OSAMPX] [--osampxpdf OSAMPXPDF]
[--minScale MINSCALE] [--seed SEED] --figure
FIGURE [--title TITLE]
[--removeFirstSamples REMOVEFIRSTSAMPLES]
[--nsampInPlot NSAMPINPLOT]
[--prettyBins PRETTYBINS]
[--logevidences LOGEVIDENCES [LOGEVIDENCES ...]]
[--coviscale COVISCALE] [--nis NIS]
[--version]
Named Arguments¶
- --version
show program’s version number and exit
Required arguments¶
- --input
Input table (tabular separated with header) with one line per cell columns with raw counts and one column nCount_RNA with total number of UMI per cell optionally other meta data to filter.
- --inputAnnData
Input annData (for example from Scanpy).
- --geneColName
Name of the column with gene counts.
- --outputs
Ouput files basename (will be npz) with different results of mcmc to combine.
- --figure
Ouput figure basename.
Optional arguments used to run MCMC¶
- --xmin
Minimum value to consider in x axis.
Default: 0
- --xmax
Maximum value to consider in x axis.
Default: 2.5
- --xscale
Possible choices: Seurat, log
scale for the x-axis: Seurat (log(1+targetSum*X)) or log (log(X))
Default: “Seurat”
- --targetSum
factor when Seurat scale is used: (log(1+targetSum*X)) (default is 10^4, use 0 for the median of nRNA_Counts)
Default: 10000
- --nx
Number of values in x to check how your evaluated pdf is compatible with the model.
Default: 100
- --osampx
Oversampling factor of x values when evaluating pdf of Poisson distribution.
Default: 10
- --osampxpdf
Oversampling factor of x values when evaluating pdf at each step of the MCMC.
Default: 5
- --minScale
Minimal value of the scale of gaussians (Default is 0.1 but cannot be smaller than max of twice the bin size of pdf evaluation and half the bin size).
Default: 0.1
- --seed
Change seed for another output.
Default: 1
Optional arguments to select input data¶
- --metadata1ColName
Name of the column with metadata1 to filter.
- --metadata1Values
Comma separated values for metadata1 of cells to keep.
- --metadata2ColName
Name of the column with metadata2 to filter.
- --metadata2Values
Comma separated values for metadata2 of cells to keep.
- --metadata3ColName
Name of the column with metadata3 to filter.
- --metadata3Values
Comma separated values for metadata3 of cells to keep.
Optional arguments to customize plots and text outputs¶
- --title
Title in figures.
- --removeFirstSamples
Number of samples to ignore before making the plots (default is nsampMCMC / 4).
- --nsampInPlot
Approximate number of samples to use in plots.
Default: 100000
- --prettyBins
Number of bins to use in plots (Default is nx).
Optional arguments to evaluate logevidence¶
- --logevidences
Ouput files of precalculated log evidence values.(if not provided will be calculated).
- --coviscale
Scale factor to apply to covariance of parameters to get random parameters in logevidence evaluation.
Default: 1
- --nis
Size of sampling of random parameters in logevidence evaluation.
Default: 1000
baredSC_2d¶
Run mcmc to get the pdf in 2D for 2 given genes using a normal distributions. The full documentation is available at https://baredsc.readthedocs.io
usage: baredSC_2d [-h] (--input INPUT | --inputAnnData INPUTANNDATA)
--geneXColName GENEXCOLNAME --geneYColName GENEYCOLNAME
[--metadata1ColName METADATA1COLNAME]
[--metadata1Values METADATA1VALUES]
[--metadata2ColName METADATA2COLNAME]
[--metadata2Values METADATA2VALUES]
[--metadata3ColName METADATA3COLNAME]
[--metadata3Values METADATA3VALUES] [--xmin XMIN]
[--xmax XMAX] [--nx NX] [--osampx OSAMPX]
[--osampxpdf OSAMPXPDF] [--minScalex MINSCALEX]
[--ymin YMIN] [--ymax YMAX] [--ny NY] [--osampy OSAMPY]
[--osampypdf OSAMPYPDF] [--minScaley MINSCALEY]
[--scalePrior SCALEPRIOR] [--scale {Seurat,log}]
[--targetSum TARGETSUM] [--nnorm NNORM]
[--nsampMCMC NSAMPMCMC] [--nsampBurnMCMC NSAMPBURNMCMC]
[--nsplitBurnMCMC NSPLITBURNMCMC] [--T0BurnMCMC T0BURNMCMC]
[--seed SEED] [--minNeff MINNEFF] [--force] --output OUTPUT
[--figure FIGURE] [--title TITLE]
[--splity SPLITY [SPLITY ...]]
[--removeFirstSamples REMOVEFIRSTSAMPLES]
[--nsampInPlot NSAMPINPLOT] [--prettyBinsx PRETTYBINSX]
[--prettyBinsy PRETTYBINSY] [--log1pColorScale]
[--logevidence LOGEVIDENCE] [--coviscale COVISCALE]
[--nis NIS] [--version]
Named Arguments¶
- --version
show program’s version number and exit
Required arguments¶
- --input
Input table (tabular separated with header) with one line per cell columns with raw counts and one column nCount_RNA with total number of UMI per cell optionally other meta data to filter.
- --inputAnnData
Input annData (for example from Scanpy).
- --geneXColName
Name of the column with gene counts for gene in x.
- --geneYColName
Name of the column with gene counts for gene in y.
- --output
Ouput file basename (will be npz) with results of mcmc.
Optional arguments to select input data¶
- --metadata1ColName
Name of the column with metadata1 to filter.
- --metadata1Values
Comma separated values for metadata1 of cells to keep.
- --metadata2ColName
Name of the column with metadata2 to filter.
- --metadata2Values
Comma separated values for metadata2 of cells to keep.
- --metadata3ColName
Name of the column with metadata3 to filter.
- --metadata3Values
Comma separated values for metadata3 of cells to keep.
Optional arguments to run MCMC¶
- --xmin
Minimum value to consider in x axis.
Default: 0
- --xmax
Maximum value to consider in x axis.
Default: 2.5
- --nx
Number of values in x to check how your evaluated pdf is compatible with the model.
Default: 50
- --osampx
Oversampling factor of x values when evaluating pdf of Poisson distribution.
Default: 10
- --osampxpdf
Oversampling factor of x values when evaluating pdf at each step of the MCMC.
Default: 4
- --minScalex
Minimal value of the scale of gaussians on x (Default is 0.1 but cannot be smaller than max of twice the bin size of pdf evaluation and half the bin size on x axis).
Default: 0.1
- --ymin
Minimum value to consider in y axis.
Default: 0
- --ymax
Maximum value to consider in y axis.
Default: 2.5
- --ny
Number of values in y to check how your evaluated pdf is compatible with the model.
Default: 50
- --osampy
Oversampling factor of y values when evaluating pdf of Poisson distribution.
Default: 10
- --osampypdf
Oversampling factor of y values when evaluating pdf at each step of the MCMC.
Default: 4
- --minScaley
Minimal value of the scale of gaussians on yx (Default is 0.1 but cannot be smaller than max of twice the bin size of pdf evaluation and half the bin size on y axis).
Default: 0.1
- --scalePrior
Scale of the truncnorm used in the prior for the correlation.
Default: 0.3
- --scale
Possible choices: Seurat, log
scale for the x-axis and y-axis: Seurat (log(1+targetSum*X)) or log (log(X))
Default: “Seurat”
- --targetSum
factor when Seurat scale is used: (log(1+targetSum*X)) (default is 10^4, use 0 for the median of nRNA_Counts)
Default: 10000
- --nnorm
Number of gaussian 2D to fit.
Default: 1
- --nsampMCMC
Number of samplings (iteractions) of mcmc.
Default: 100000
- --nsampBurnMCMC
Number of samplings (iteractions) in the burning phase of mcmc (Default is nsampMCMC / 4).
- --nsplitBurnMCMC
Number of steps in the burning phase of mcmc.
Default: 10
- --T0BurnMCMC
Initial temperature in the burning phase of mcmc.
Default: 100.0
- --seed
Change seed for another output.
Default: 1
- --minNeff
Will redo the MCMC with 10 times more samples until the number of effective samples that this value (Default is not set so will not rerun MCMC).
- --force
Force to redo the mcmc even if output exists.
Optional arguments to get plots and text outputs¶
- --figure
Ouput figure basename.
- --title
Title in figures.
- --splity
Threshold value to plot the density for genex for 2 categories in geney values.
- --removeFirstSamples
Number of samples to ignore before making the plots (default is nsampMCMC / 4).
- --nsampInPlot
Approximate number of samples to use in plots.
Default: 100000
- --prettyBinsx
Number of bins to use in x in plots (Default is nx).
- --prettyBinsy
Number of bins to use in y in plots (Default is ny).
- --log1pColorScale
Use log1p color scale instead of linear color scale.
Default: False
Optional arguments to get logevidence¶
- --logevidence
Ouput file to put logevidence value.
- --coviscale
Scale factor to appy to covariance of parameters to get random parameters in logevidence evaluation.
Default: 1
- --nis
Size of sampling of random parameters in logevidence evaluation.
Default: 1000
combineMultipleModels_2d¶
Combine mcmc 2D results from multiple models to get a mixture using logevidence to infer weights.
usage: combineMultipleModels_2d [-h]
(--input INPUT | --inputAnnData INPUTANNDATA)
--geneXColName GENEXCOLNAME --geneYColName
GENEYCOLNAME
[--metadata1ColName METADATA1COLNAME]
[--metadata1Values METADATA1VALUES]
[--metadata2ColName METADATA2COLNAME]
[--metadata2Values METADATA2VALUES]
[--metadata3ColName METADATA3COLNAME]
[--metadata3Values METADATA3VALUES] --outputs
OUTPUTS [OUTPUTS ...] [--xmin XMIN]
[--xmax XMAX] [--nx NX] [--osampx OSAMPX]
[--osampxpdf OSAMPXPDF]
[--minScalex MINSCALEX] [--ymin YMIN]
[--ymax YMAX] [--ny NY] [--osampy OSAMPY]
[--osampypdf OSAMPYPDF]
[--minScaley MINSCALEY] [--scale {Seurat,log}]
[--scalePrior SCALEPRIOR]
[--targetSum TARGETSUM] [--seed SEED] --figure
FIGURE [--title TITLE]
[--splity SPLITY [SPLITY ...]]
[--removeFirstSamples REMOVEFIRSTSAMPLES]
[--nsampInPlot NSAMPINPLOT]
[--prettyBins PRETTYBINS]
[--prettyBinsx PRETTYBINSX]
[--prettyBinsy PRETTYBINSY]
[--log1pColorScale] [--getPVal]
[--logevidences LOGEVIDENCES [LOGEVIDENCES ...]]
[--coviscale COVISCALE] [--nis NIS]
[--version]
Named Arguments¶
- --version
show program’s version number and exit
Required arguments¶
- --input
Input table (tabular separated with header) with one line per cell columns with raw counts and one column nCount_RNA with total number of UMI per cell optionally other meta data to filter.
- --inputAnnData
Input annData (for example from Scanpy).
- --geneXColName
Name of the column with gene counts for gene in x.
- --geneYColName
Name of the column with gene counts for gene in y.
- --outputs
Ouput files basename (will be npz) with different results of mcmc to combine.
- --figure
Ouput figure basename.
Optional arguments used to run MCMC¶
- --xmin
Minimum value to consider in x axis.
Default: 0
- --xmax
Maximum value to consider in x axis.
Default: 2.5
- --nx
Number of values in x to check how your evaluated pdf is compatible with the model.
Default: 50
- --osampx
Oversampling factor of x values when evaluating pdf of Poisson distribution.
Default: 10
- --osampxpdf
Oversampling factor of x values when evaluating pdf at each step of the MCMC.
Default: 4
- --minScalex
Minimal value of the scale of gaussians on x (Default is 0.1 but cannot be smaller than max of twice the bin size of pdf evaluation and half the bin size on x axis).
Default: 0.1
- --ymin
Minimum value to consider in y axis.
Default: 0
- --ymax
Maximum value to consider in y axis.
Default: 2.5
- --ny
Number of values in y to check how your evaluated pdf is compatible with the model.
Default: 50
- --osampy
Oversampling factor of y values when evaluating pdf of Poisson distribution.
Default: 10
- --osampypdf
Oversampling factor of y values when evaluating pdf at each step of the MCMC.
Default: 4
- --minScaley
Minimal value of the scale of gaussians on yx (Default is 0.1 but cannot be smaller than max of twice the bin size of pdf evaluation and half the bin size on y axis).
Default: 0.1
- --scale
Possible choices: Seurat, log
scale for the x-axis and y-axis: Seurat (log(1+targetSum*X)) or log (log(X))
Default: “Seurat”
- --scalePrior
Scale of the truncnorm used in the prior for the correlation.
Default: 0.3
- --targetSum
factor when Seurat scale is used: (log(1+targetSum*X)) (default is 10^4, use 0 for the median of nRNA_Counts)
Default: 10000
- --seed
Change seed for another output.
Default: 1
Optional arguments to select input data¶
- --metadata1ColName
Name of the column with metadata1 to filter.
- --metadata1Values
Comma separated values for metadata1 of cells to keep.
- --metadata2ColName
Name of the column with metadata2 to filter.
- --metadata2Values
Comma separated values for metadata2 of cells to keep.
- --metadata3ColName
Name of the column with metadata3 to filter.
- --metadata3Values
Comma separated values for metadata3 of cells to keep.
Optional arguments to customize plots and text outputs¶
- --title
Title in figures.
- --splity
Threshold value to plot the density for genex for 2 categories in geney values.
- --removeFirstSamples
Number of samples to ignore before making the plots (default is nsampMCMC / 4).
- --nsampInPlot
Approximate number of samples to use in plots.
Default: 100000
- --prettyBins
Number of bins to use in plots (Default is nx).
- --prettyBinsx
Number of bins to use in x in plots (Default is nx).
- --prettyBinsy
Number of bins to use in y in plots (Default is ny).
- --log1pColorScale
Use log1p color scale instead of linear color scale.
Default: False
- --getPVal
Use less samples to get an estimation of the p-value.
Default: False
Optional arguments to evaluate logevidence¶
- --logevidences
Ouput files of precalculated log evidence values.(if not provided will be calculated).
- --coviscale
Scale factor to apply to covariance of parameters to get random parameters in logevidence evaluation.
Default: 1
- --nis
Size of sampling of random parameters in logevidence evaluation.
Default: 1000