Basic Usage

This section shows how to use GeneCover to compute gene-gene correlation and select marker genes.

Required Imports

from genecover import *
import numpy as np

data: A NumPy array of shape (N, d) storing gene expression measurements, where N is the number of cells/spots and d is the number of genes.

Alternatively, data can be a list of such arrays (e.g., one per donor or batch), as long as they all have the same number of genes.
w: A 1D array of length d representing the cost (weight) of selecting each gene. A simple choice is:
```
w = np.ones(data.shape[1])
```
solver: A string indicating which solver to use for gene selection. Options include:
- “Gurobi”: the Gurobi solver for integer programming.
- “SCIP”: the SCIP solver for integer programming.
- “Greedy”: the greedy heuristic for gene selection.

Compute the gene-gene correlation matrix:

corr_mat = gene_gene_correlation(data)

Obtain 100 marker genes using the standard Gurobi optimization:

markers = GeneCover(
    num_marker=100,
    corr_mat=corr_mat,
    w=w,
    solver=solver)

Select 300 markers in three iterative steps (e.g., 100 markers per round):

iterative_markers = Iterative_GeneCover(
    incremental_sizes=[100, 100, 100],
    corr_mat=corr_mat,
    w=w,
    solver=solver)