Automated Methods for the Analysis of Skeletal Muscle Fiber Size and Metabolic Type - International Review of Cell and Molecular Biology - page 312

4), a protein from this ontology group, has been previously linked to with-
drawal of skeletal muscle myoblasts from the cell cycle ( Tomczak et al.,
2004 ). Also intriguingly, five proteins identified using this approach
(EGFR, ERBB2 (aka HER2/neu), MYC, MAPK3, and BRAF) are known
members of the “endometrial cancer pathway” identified using the Kyoto
Encyclopedia of Genes and Genomes (KEGG) gene enrichment option
( Fig. 7.16 D). While this similarity is initially surprising, considering that
skeletal muscle does not typically undergo cell division, skeletal muscle does
undergo “hypertrophic growth” (increase in skeletal muscle fiber size) dur-
ing which expression of c-MYC, a proto-oncogene, is increased ( von
Walden et al., 2012 ). Relatedly, proto-oncogenes are intimately involved
in the hypertrophic growth of cardiac muscle ( Gelb and Tartaglia, 2011;
Oceandy et al., 2009 ), a tissue type with many similarities to skeletal muscle.
We envision that this histospatial phenotypic approach using HPA
images, combined with the bioinformatic analysis of the resulting PPI net-
works, will lead to a multitude of testable hypotheses regarding protein
expression and function in human skeletal muscle. An added advantage of
the HPA is that it is entirely based upon human tissue, and the open-source
images are available through the Internet to interested researchers. A further
advantage of a bioinformatics approach is identifying orthologous proteins in
other model organisms to develop inferences regarding possible protein
functionality in humans. A successful application of this approach has been
the discovery and elucidation of the MAPK and stress kinase cascades, which
are conserved from yeast to man and which participate in pathologies rang-
ing from obesity, liver function, and cancer ( Johnson and Lapadat, 2002 ).
We envision that this histospatial phenotypic method can be vastly improved
for speed and accuracy, by the development of automated image-analysis
algorithms, a subject area that will likely represent a very fruitful area of col-
laboration between computational scientists and biologists. Furthermore, it
is expected that histospatial phenotyping can be adapted and optimized for
other tissues with obvious substructures, such as glands, epithelial/endothe-
lial sheets, and blood vessels, or distinction between tumor and stroma.
8. SUMMARY
The exquisitely patterned organization of skeletal muscle, in which
there are obvious correlations between the organization of the proteins,
fibers, and contractile function, makes this tissue type an ideal subject for
digital imaging and HCA strategies. In this presentation, we have reviewed