Variable-Selection Emerges on Top in Empirical Comparison of Whole-Genome Complex-Trait Prediction Methods

Accurate prediction of complex traits based on whole-genome data is a computational problem of paramount importance, particularly to plant and animal breeders. However, the number of genetic markers is typically orders of magnitude larger than the number of samples (p >> n), amongst other challenges. We assessed the effectiveness of a diverse set of state-of-the-art methods on publicly accessible real data. The most surprising finding was that approaches with feature selection performed better than others on average, in contrast to the expectation in the community that variable selection is mostly ineffective, i.e. that it does not improve accuracy of prediction, in spite of p >> n. We observed superior performance despite a somewhat simplistic approach to variable selection, possibly suggesting an inherent robustness. This bodes well in general since the variable selection methods usually improve interpretability without loss of prediction power. Apart from identifying a set of benchmark data sets (including one simulated data), we also discuss the performance analysis for each data set in terms of the input characteristics.     

Guar gum derivatives. Part I: Preparation and properties

O-(2-Hydroxyethyl), O-(2-hydroxypropyl) and O-carboxymethyl derivatives of guar gum have been prepared under different experimental conditions. Several properties such as moisture regain, rate of hydration, solubility, viscosity and rheology of these derivatives have been studied. The properties depend upon polysaccharide chain length, and the nature and degree of chemical modification. The effect of alkali and alkaline hydrogen peroxide on the properties of guar gum have also been studied.     

Smad3 signaling is required for satellite cell function and myogenic differentiation of myoblasts

TGF-β and myostatin are the two most important regulators of muscle growth. Both growth factors have been shown to signal through a Smad3-dependent pathway. However to date, the role of Smad3 in muscle growth and differentiation is not investigated. Here, we demonstrate that Smad3-null mice have decreased muscle mass and pronounced skeletal muscle atrophy. Consistent with this, we also find increased protein ubiquitination and elevated levels of the ubiquitin E3 ligase MuRF1 in muscle tissue isolated from Smad3-null mice. Loss of Smad3 also led to defective satellite cell (SC) functionality. Smad3-null SCs showed reduced propensity for self-renewal, which may lead to a progressive loss of SC number. Indeed, decreased SC number was observed in skeletal muscle from Smad3-null mice showing signs of severe muscle wasting. Further in vitro analysis of primary myoblast cultures identified that Smad3-null myoblasts exhibit impaired proliferation, differentiation and fusion, resulting in the formation of atrophied myotubes. A search for the molecular mechanism revealed that loss of Smad3 results in increased myostatin expression in Smad3-null muscle and myoblasts. Given that myostatin is a negative regulator, we hypothesize that increased myostatin levels are responsible for the atrophic phenotype in Smad3-null mice. Consistent with this theory, inactivation of myostatin in Smad3-null mice rescues the muscle atrophy phenotype.     

Decorin binds myostatin and modulates its activity to muscle cells

Myostatin, a member of TGF-beta superfamily of growth factors, acts as a negative regulator of skeletal muscle mass. The mechanism whereby myostatin controls the proliferation and differentiation of myogenic cells is mostly clarified. However, the regulation of myostatin activity to myogenic cells after its secretion in the extracellular matrix (ECM) is still unknown. Decorin, a small leucine-rich proteoglycan, binds TGF-beta and regulates its activity in the ECM. Thus, we hypothesized that decorin could also bind to myostatin and participate in modulation of its activity to myogenic cells. In order to test the hypothesis, we investigated the interaction between myostatin and decorin by surface plasmon assay. Decorin interacted with mature myostatin in the presence of concentrations of Zn(2+) greater than 10microM, but not in the absence of Zn(2+). Kinetic analysis with a 1:1 binding model resulted in dissociation constants (K(D)) of 2.02x10(-8)M and 9.36x10(-9)M for decorin and the core protein of decorin, respectively. Removal of the glycosaminoglycan chain by chondroitinase ABC digestion did not affect binding, suggesting that decorin could bind to myostatin with its core protein. Furthermore, we demonstrated that immobilized decorin could rescue the inhibitory effect of myostatin on myoblast proliferation in vitro. These results suggest that decorin could trap myostatin and modulate its activity to myogenic cells in the ECM.     

Myostatin induces cachexia by activating the ubiquitin proteolytic system through an NF-kappaB-independent, FoxO1-dependent mechanism

Myostatin, a transforming growth factor-beta (TGF-beta) super-family member, has been well characterized as a negative regulator of muscle growth and development. Myostatin has been implicated in several forms of muscle wasting including the severe cachexia observed as a result of conditions such as AIDS and liver cirrhosis. Here we show that Myostatin induces cachexia by a mechanism independent of NF-kappaB. Myostatin treatment resulted in a reduction in both myotube number and size in vitro, as well as a loss in body mass in vivo. Furthermore, the expression of the myogenic genes myoD and pax3 was reduced, while NF-kappaB (the p65 subunit) localization and expression remained unchanged. In addition, promoter analysis has confirmed Myostatin inhibition of myoD and pax3. An increase in the expression of genes involved in ubiquitin-mediated proteolysis is observed during many forms of muscle wasting. Hence we analyzed the effect of Myostatin treatment on proteolytic gene expression. The ubiquitin associated genes atrogin-1, MuRF-1, and E214k were upregulated following Myostatin treatment. We analyzed how Myostatin may be signaling to induce cachexia. Myostatin signaling reversed the IGF-1/PI3K/AKT hypertrophy pathway by inhibiting AKT phosphorylation thereby increasing the levels of active FoxO1, allowing for increased expression of atrophy-related genes. Therefore, our results suggest that Myostatin induces cachexia through an NF-kappaB-independent mechanism. Furthermore, increased Myostatin levels appear to antagonize hypertrophy signaling through regulation of the AKT-FoxO1 pathway.     

Myostatin-deficient mice lose more skeletal muscle mass than wild-type controls during hindlimb suspension

Myostatin inhibits myogenesis. Therefore, we sought to determine if mice lacking the myostatin gene [Mstn(-/-)] would lose less muscle mass than wild-type mice during 7 days of hindlimb suspension (HS). Male Mstn(-/-) and wild-type (C57) mice were subjected to HS or served as ground-based controls (n = 6/group). Wild-type mice lost 8% of body mass and approximately 13% of wet mass from biceps femoris, quadriceps femoris, and soleus, whereas the mass of extensor digitorum longus (EDL) was unchanged after HS. Unexpectedly, Mstn(-/-) mice lost more body (13%, P < 0.05) and quadriceps femoris (17%, P < 0.05) mass than wild-type mice and lost 33% of EDL mass (P < 0.01) after HS. Protein expression of myostatin in biceps femoris and quadriceps femoris was not altered, whereas expression of MyoD, Myf-5, and myogenin increased in wild-type mice and tended to decrease in muscles of Mstn(-/-) mice. These data suggest that HS induced myogenesis in wild-type mice to counter atrophy, whereas myogenesis was not induced in Mstn(-/-) mice, thereby resulting in a greater loss of muscle mass.     

Molecular expression of myostatin and MyoD is greater in double-muscled than normal-muscled cattle fetuses

Excessive muscling in double-muscled cattle arises from mutations in the myostatin gene, but the role of myostatin in normal muscle development is unclear. The aim of this study was to measure the temporal relationship of myostatin and myogenic regulatory factors during muscle development in normal (NM)- and double-muscled (DM) cattle to determine the timing and possible targets of myostatin action in vivo. Myostatin mRNA peaked at the onset of secondary fiber formation (P < 0.001) and was greater in DM (P < 0.001) than in NM. MyoD expression was also elevated throughout primary and secondary fiber formation (P < 0.001) and greater in DM (P < 0.05). Expression of myogenin peaked later than MyoD (P < 0.05); however, it did not differ between NM and DM. These data show that myostatin and MyoD increase coincidentally during formation of muscle fibers, indicating a coordinated role in the terminal differentiation and/or fusion of myoblasts. Myostatin mRNA is also consistently higher in DM than NM, suggesting that a feedback loop of regulation is also disrupted in the myostatin-deficient condition.