Workflow

Here the different analysis pipelines, when using pymportx, are described.

Previous Packages

---

As mentioned before, the pymportx package is compatible for fast gene count estimation with the following upstream trasncript quantification methods: Salmon, Sailfish, Kallisto, and RSEM. The package employment for each method is described in the Usage Tutorial.

Downstream Analysis Packages

---

PyDESeq2 and decoupleR

The anndata output from pymportx can be used to perform a Differential Expression Analysis using PyDESeq2. Other analyses, such as reads quality control or volcano plotting, can be also conducted using the decoupleR package. Click on the links to access their documentation for more details.

Here is an example for Differential Expression Analysis using PyDESeq2 taken from decoupleR documentation:

# Import PyDESeq2
from pydeseq2.dds import DeseqDataSet, DefaultInference
from pydeseq2.ds import DeseqStats

# Build PyDESeq2 object
inference = DefaultInference(n_cpus=8)
dds = DeseqDataSet(adata=adata,
    design_factors='condition',
    refit_cooks=True,
    inference=inference)

# Compute LFCs
dds.deseq2()

# Extract contrast between COVID-19 vs normal
stat_res = DeseqStats(dds,
    contrast=["condition", 'treatment', 'control'],
    inference=inference)

# Compute Wald test
stat_res.summary()

# Extract results
results_df = stat_res.results_df
results_df

PyWGCNA

The following example demonstrates how to perform a gene co-expression network analysis using PyWGCNA:

wgcna = PyWGCNA.WGCNA(anndata=adata_cleaned)

wgcna.preprocess()

wgcna.findModules()

wgcna.analyseWGCNA()