UC San Diego

The human muscular system represents nearly 75% of the body mass and encompasses two major muscle forms- striated and smooth. Striated muscle, composed broadly of myofibers, accompanying membrane systems, cytoskeletal networks together with the metabolic and regulatory machinery, have revealed complexities in composition, structure and function. A disruption to any component within this complex

system of interactions lead to disorders of the muscle, typically characterized by muscle fiber loss, reduced motor output and in some cases death. Advent of high-throughput technologies coupled with elegant approaches to deciphering data using bioinformatics and systems biology, are providing new venues for detailed exploration of mammalian muscle.

This dissertation describes the use of publicly available high-throughput data, in conjunction with co-expression network methodologies developed for a comprehensive, interpretable systems-level perspective on mechanisms underlying associated muscle pathologies. This study begins with the exploration of the temporal transcriptional response of skeletal muscle to Botulinum Neurotoxin-A (Botox ®) over a 1-year period, in the framework of muscle physiology. Next, utilizing co-expression network analysis, putative markers associated with recovery of muscle trophicity are identified, furthermore providing an unbiased validation of the response documented earlier. These studies represent the first attempt at categorically assessing the whole-transcriptomic changes associated with BoNT-A treatment in muscle.

The latter half of this research focuses on discerning patho-mechanisms of human diseases affecting muscle. Particularly, co-expression network statistics are leveraged to identify dysregulated pathways and biomarkers of disease progression, underlying duchenne muscular dystrophy. Next, a quantitative framework integrating transcriptional, protein interaction, and drug-target data is developed to extract functional similarities and mechanisms amongst 20 diseases affecting the muscle. Lastly, an approach to differential co-expression analysis using signed and weighted co-expression networks is described. This approach is subsequently utilized to assess and identify differential mechanisms underlying ischemic and idiopathic dilated cardiomyopathy. The analysis and results from the aforementioned studies have enabled a deeper understanding of the complex interactions underlying muscle pathologies; providing opportunities for drug development and personalized medicine.