The cost affects the size of the inferred sub-network. Increasing the cost, decreases the size of the inferred sub-network but as the algorithm tends to find the most ‘explanatory edges’ first, irrespective of the cost parameter smaller networks are mostly contained within the larger ones.Using the Cost Values
The edge cost sweep is an optimization value. In case of fewer solutions or sub-optimal ranking (e.g High number of genes sharing the same rank) it will require additional tuning by yourself. Sometimes several runs with different parameters and settings will be needed.
The Step Size is sensitive, when using a small value generally generate more interesting rankings. However, the the runtime will increase drastically.