We have utilized hidden Markov models using HMMER software to predict and generate putative strong secretory signal peptide sequences for directing efficient secretion of cytokines from skeletal muscle for therapeutic applications. The results show that this approach can analyze signal sequences of a skeletal muscle secretome dataset and classify them, emitting new sequences that are strong candidate skeletal muscle-enriched signal peptides. The emitted signal peptides also were analyzed for their hydropathy and secondary structure profiles as compared to native signal peptides. The emitted signal peptides had a higher degree of hydropathy and helical composition relative to native sequences, which may suggest that these new sequences may hold promize for promoting enhanced secretion of proteins including cytokines or propeptides from skeletal muscle.