Marker gene detection

library(Giotto)
data("mini_giotto_single_cell")

This tutorial starts from a pre-computed mini Giotto object, which has already undergone pre-processing, dimensions reduction and clustering steps. Currently provides 3 different methods to identify marker genes:
- using a new Gini-index method
- using Scran
- using Mast

Each method can either identify marker genes between 2 selected (groups of) clusters or for each individual cluster.

\(~\)

1. Gini
# between 2 groups
gini_markers = findGiniMarkers(gobject = mini_giotto_single_cell,
                               cluster_column = 'leiden_clus',
                               group_1 = 1,
                               group_2 = 2)
# for each cluster
gini_markers = findGiniMarkers_one_vs_all(gobject = mini_giotto_single_cell,
                                          cluster_column = 'leiden_clus')

\(~\)

2. Scran

Requires Scran to be installed.

# between 2 groups
scran_markers = findScranMarkers(gobject = mini_giotto_single_cell,
                                 cluster_column = 'leiden_clus',
                                 group_1 = 1,
                                 group_2 = 2)
# for each cluster
scran_markers = findScranMarkers_one_vs_all(gobject = mini_giotto_single_cell,
                                            cluster_column = 'leiden_clus')

\(~\)

3. Mast

Requires Mast to be installed.

# between 2 groups
mast_markers = findMastMarkers(gobject = mini_giotto_single_cell,
                                cluster_column = 'leiden_clus',
                                group_1 = 1,
                                group_2 = 2)

# for each cluster
mast_markers = findMastMarkers_one_vs_all(gobject = mini_giotto_single_cell,
                                          cluster_column = 'leiden_clus')

\(~\)

4. Wrapper

A general wrapper has also been created which covers all three methods:
see findMarkers and findMarkers_one_vs_all and specify the method parameter.