Description Usage Arguments Details Author(s)
This function extracts transpriptomic programs
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | scandal_programs_of_intra_sample_heterogeneity(
object,
samples = NULL,
clustering_data = NULL,
algorithm = "nmf",
rank = 10,
ngenes1 = 50,
ngenes2 = 30,
sd_threshold = 0.8,
filter_method = "relative",
n_clusters_min = 2,
n_clusters_max = 10,
bin_control = TRUE,
n_control_bins = 25,
n_bin_genes = 100,
return_all = FALSE,
verbose = FALSE,
...
)
|
... |
further arguments passed to nmf function |
Generally, the algorithm extracts programs in a bottoms-up approach
starting from extracting programs within each individual sample. The algorithm
then detecs highly variable programs that best represent coherent program clusters
and aggregates these clusters into programs that generalize accross all samples.
The algorithm performs the following steps:
Call prepare_samples to prepare a ScandalDataSet for
each individual sample. This step can be bypassed by supplying a valid samples
argument.
For each individual sample generate a predefined number of clusters
(configurable by the rank
parameter). At the moment the only supported
clustering algorithm is NMF via nmf_run. This step can be bypassed by supplying
a valid clustering_data
argument.
For each sample extract an initial list of within-sample programs by calling
nmf_extract_programs. The number of genes in each program is configurable by
the ngenes1
parameter. The number of initial within-sample programs correpsonds
to rank
.
For each sample score the cells for each of the within-sample programs by calling score_within_samples. See scalop::score for more details about how to score cells while controling for differences in cell complexities.
For each sample compute the standard deviation of scores for each within-sample program by calling compute_programs_sd.
Avishay Spitzer
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.