DeepFrag-k: a fragment-based deep learning approach for protein fold recognition.

Background: One of the most essential problems in structural bioinformatics is protein fold recognition. In this paper, we design a novel deep learning architecture, so-called DeepFrag-k, which identifies fold discriminative features at fragment level to improve the accuracy of protein fold recognition. DeepFrag-k is composed of two stages: the first stage employs a multi-modal Deep Belief Network (DBN) to predict the potential structural fragments given a sequence, represented as a fragment vector, and then the second stage uses a deep convolutional neural network (CNN) to classify the fragment vector into the corresponding fold.

Dormant pathogenic CD4 T cells are prevalent in the peripheral repertoire of healthy mice.

Thymic central tolerance eliminates most immature T cells with autoreactive T cell receptors (TCR) that recognize self MHC/peptide complexes. Regardless, an unknown number of autoreactive CD4Foxp3 T cells escape negative selection and in the periphery require continuous suppression by CD4Foxp3 regulatory cells (Tregs). Here, we compare immune repertoires of Treg-deficient and Treg-sufficient mice to find Tregs continuously constraining one-third of mature CD4Foxp3 cells from converting to pathogenic effectors in healthy mice.

Non-canonicaly recruited TCRαβCD8αα IELs recognize microbial antigens.

In the gut, various subsets of intraepithelial T cells (IELs) respond to self or non-self-antigens derived from the body, diet, commensal and pathogenic microbiota. Dominant subset of IELs in the small intestine are TCRαβCD8αα cells, which are derived from immature thymocytes that express self-reactive TCRs. Although most of TCRαβCD8αα IELs are thymus-derived, their repertoire adapts to microbial flora.

Conformational Clusters of Phosphorylated Tyrosine.

Tyrosine phosphorylation plays an important role in many cellular and intercellular processes including signal transduction, subcellular localization, and regulation of enzymatic activity. In 1999, Blom et al., using the limited number of protein data bank (PDB) structures available at that time, reported that the side chain structures of phosphorylated tyrosine (pY) are partitioned into two conserved conformational clusters ( Blom, N.

ICOSA: A Distance-Dependent, Orientation-Specific Coarse-Grained Contact Potential for Protein Structure Modeling.

The relative distance and orientation in contacting residue pairs plays a significant role in protein folding and stabilization. We hereby devise a new knowledge-based, coarse-grained contact potential, so-called ICOSA, by correlating inter-residue contact distance and orientation in evaluating pair-wise inter-residue interactions. The rationale of our approach is to establish icosahedral local coordinates to estimate the statistical residue contact distributions in all spherical triangular shells within a sphere.