Computational Systems Biology Methods and Protocols.7z

MAST

A logistic regression model is

used to test differentialexpression rate betweengroups while a Gaussiangeneralized linear model(GLM) describes expressionconditionally on non-zeroexpression estimates. Modelsare corrected for cellulardetection rate

Requires normalized gene

expression estimates andprovides gene-specific P valuesfrom summing likelihood ratioor Wald tests from the twocomponents

Framework Command

line

Unix/

Linux,Mac OS,Windows

R[

90

]

D3E

A tool for identifying

differentially expressed genes,based on single-cell RNA-seqdata. D3E consists of twomodules: one for identifyingdifferentially expressed(DE) genes and one for fittingthe parameters of a Poisson-Beta distribution

D3E takes a read count table as

an input, with rows andcolumns corresponding totranscripts and cells,respectively. The user shouldsplit the columns into two ormore groups by providing celllabels in the input file. If thereare more than two groups ofcells, they must be comparedone pair at a time

Package

Command

line

Unix/

Linux

Python

[^91

]

scDD

Models expressed counts as a

Dirichlet process mixture(DPM) of normals to test fordifferentially distributed(DD) genes associated withmultimodality in the expressedcomponent. Samples from theposterior further characterizethe gene-specificdistributional differencebetween two biologicalconditions to identify genesthat are differentiallyexpressed (DE), differ in the

Requires normalized,

log-transformed geneexpression estimates andprovides gene-specific P values(or a false discovery rate(FDR)-controlled list) of DDgenes between two biologicalconditions. Each DD gene isthen classified into a specifictype of distributionaldifference

Framework,

package

Command

line

Unix/

Linux,Mac OS,Windows

R[

92

,^93

]

(continued)

Applications of Single-Cell Sequencing for Multiomics 341