Step 6 — K-means Clustering

Partitions genes into k co-regulated groups based on their expression profiles across samples / conditions.

Inputs

• The expression matrix from Step 4 / 5.

Settings

• k — number of clusters. Use the elbow + silhouette plot to pick.

• Max iterations — default 300.

• Random seed — deterministic by default (seed = 42), exposed for reproducibility.

Diagnostics

• Elbow plot — within-cluster SSE vs k for k = 2 … 15.

• Silhouette plot — mean silhouette score vs k.

Outputs

• clusters/cluster_1.txt … cluster_k.txt — gene lists per cluster.

• cluster_centers.png — average expression profile per cluster.