Graph algorithms can be effective tools for analyzing the immense data sets that frequently arise from high-throughput biological experiments. The current study has two parts. First, new clique-centric methods are applied to the notoriously difficult problem of combining transcriptomic and proteomic data. This is done by anchoring cliques on specific proteomic targets in order to find networks of transcripts related to the protein. Second, the paraclique algorithm is applied to transcriptomic data under various parameters. Non-obese diabetic mice are used as a target organism. Proteomic results are compared to a list of transcripts generated by rank correlation and found to produce networks more inclusive of the protein target. Paraclique results are compared with results from K-means clustering. The former finds better networks according to both biological and statistical metrics.