Adding spearman correlation to findmodules

[Buiboni]

Spearman correlation is widely use in biology especially when dealing with semi-quantitative data. This pull request implement spearman correlation in the main function findmodules.

[nargesr]

Thank you @Buiboni

It's wonderful. Could you add a docstring describing the spearman parameter you used for those functions?

Once you add that information, I can start merging your pull request.

Best,
Narges

[gitbenlewis]

@Buiboni here is a docstring for your pull request (appended to class doc string attributes section)

class WGCNA(GeneExp):
"""
A class used to create WGCNA (weighted gene co-expression network analysis).
Attributes
----------
name: str
name of the WGCNA we used to visualize data (default: 'WGCNA')
geneExpression : class GeneExp
gene expression class
geneExpPath: str
geneExp: dataframe
sep: str
datExpr: anndata
clean gene expression matrix with genes and sample information
TPMcutoff: int
cut off for removing genes with out any specific expression
power : int
power used to have scale free network (default: 6)
networkType : str
Type of network we can create (default: "signed hybrid")
possible Type: "unsigned", "signed" and "signed hybrid"
TOMType: str
Type of topological overlap matrix(TOM) (default: "signed")
possible Type: "unsigned", "signed"
minModuleSize : int
We like large modules, so we set the minimum module size relatively high (default: 50)
naColor: str
color we used to identify genes we don't find any cluster for them (default: "grey")
cut: float
number to remove outlier sample (default: 'inf')
By default we don't remove any sample by hierarchical clustering
MEDissThres: int
(default: 0.2)
save: bool
define wether you want to save result of important steps or not (default: False)
outputPath: str
if save is TRUE, define where you want to save your data (default: '')
    spearman : bool, default False
        If True, compute correlations using Spearman correlation instead of the default Pearson correlation. This is achieved by replacing the expression data with their ranks before computing the correlation.