arboreto-grn-inference

To infer gene regulatory networks from expression data, provide an expression matrix and optionally a transcription factor list to the Skill, which uses GRNBoost2 or GENIE3 algorithms to identify regulatory relationships.

Yes, you can scale GRN inference for large single-cell RNA-seq datasets using Dask parallelization, which distributes the computational workload of the network construction algorithms across multiple cores or workers.

Both algorithms infer transcription factor-target gene relationships from expression data, but GRNBoost2 typically offers faster gradient boosting performance, while GENIE3 uses tree-based methods, with both scaling efficiently through Dask integration.

A transcription factor list is optional but recommended for accurate gene regulatory network construction, as it constrains the algorithms to specifically evaluate known transcription factors rather than testing all genes as potential regulators.

The Skill processes gene expression data using pandas, numpy, and anndata structures, requiring formatted expression matrices like TSV files to identify transcription factor-target relationships through GRNBoost2 or GENIE3 algorithms.

After inferring the gene regulatory network, you can visualize the resulting transcription factor-target gene relationships by processing the network output with networkx for graph structure and matplotlib for graphical rendering.