Two papers in PLoS Computational Biology this week try to better understand the three-dimensional structure of proteins. In one, scientists at the Max Planck Institute for Molecular Genetics developed a method to more precisely find the minimal subset of residues that can define the global structure of a protein. Their strategy, called "cone-peeling," outperformed using random sets of residues and will open "new avenues in the fields of structure prediction, empirical potentials, and docking," they say. In the second paper, researchers led by Mona Singh and Thomas Funkhouser at Princeton University present ConCavity, a ligand binding site prediction algorithm for proteins that "integrates evolutionary sequence conservation estimates with structure-based methods for identifying protein surface cavities."
A paper in PLoS Genetics presents results from the Stockholm Atherosclerosis Gene Expression study. In looking at gene expression changes associated with developing coronary artery disease across liver, skeletal muscle, visceral fat, and arterial wall tissue, the researchers found a signature consisting of 128 genes that increased risk for CAD. This atherosclerosis module involves transendothelial migration of leukocytes and LIM domain binding 2 as its high-hierarchy regulator, they say. "Our study design represents a novel way of understanding the molecular underpinnings of CAD, focusing on genome-wide expression sensing both environmental and genetic influences," the author summary says.
Research led by Avinash Abhyankar at Lund University looked at non-protein coding mitochondrial DNA in rats. Comparing 27 Rattus norvegicus mtDNA sequences, the group found two variable positions in 12S rRNA, 20 in 16S rRNA, eight within the tRNA genes, and 13 in the D-loop. Based on their analysis of conserved sequences, they propose that some of the variability within the 16S rRNA, tRNA-Cys, and the D-loop might be important for mitochondrial function and its regulation. Their work was published this week in PLoS One.
In other work in PLoS One, scientists take a proteomic view of the Staphylococcus aureus pathogen. Researchers from the Ernst-Moritz-Arndt-University Greifswald and the University of Groningen used a quantitative proteomics approach to identify 1,700 proteins and quantify 1,450 of them across four subproteomic fractions: cytosolic proteins, membrane-bound proteins, cell surface-associated proteins, and extracellular proteins, which covers about three-fourths of the entire proteome. "Our model study represents the most comprehensive quantification of a bacterial proteome reported to date," they say in the abstract.