Mannose-binding lectin gene polymorphism predicts hospital admissions for COPD infections

Infection frequently causes exacerbations of chronic obstructive pulmonary disease (COPD). Mannose-binding lectin (MBL) is a pattern-recognition receptor that assists in clearing microorganisms. Polymorphisms in the MBL2 gene reduce serum MBL levels and are associated with risk of infection. We studied whether the MBL2 codon 54 B allele affected serum MBL levels, admissions for infective exacerbation in COPD and disease susceptibility. Polymorphism frequency was determined by PCR-RFLP in 200 COPD patients and 104 smokers with normal lung function. Serum MBL was measured as mannan-binding activity in a subgroup of 82 stable COPD patients. Frequency of COPD admissions for infective exacerbation was ascertained for a 2-year period. The MBL2 codon 54 B allele reduced serum MBL in COPD patients. In keeping, patients carrying the low MBL-producing B allele had increased risk of admission for infective exacerbation (OR 4.9, P-corrected = 0.011). No association of MBL2 genotype with susceptibility to COPD was detected. In COPD, serum MBL is regulated by polymorphism at codon 54 in its encoding gene. Low MBL-producing genotypes were associated with more frequent admissions to hospital with respiratory infection, suggesting that the MBL2 gene is disease-modifying in COPD. MBL2 genotype should be explored prospectively as a prognostic marker for infection risk in COPD.

The applicability of recurrent neural networks for biological sequence analysis

Selection of machine learning techniques requires a certain sensitivity to the requirements of the problem. In particular, the problem can be made more tractable by deliberately using algorithms that are biased toward solutions of the requisite kind. In this paper, we argue that recurrent neural networks have a natural bias toward a problem domain of which biological sequence analysis tasks are a subset. We use experiments with synthetic data to illustrate this bias. We then demonstrate that this bias can be exploitable using a data set of protein sequences containing several classes of subcellular localization targeting peptides. The results show that, compared with feed forward, recurrent neural networks will generally perform better on sequence analysis tasks. Furthermore, as the patterns within the sequence become more ambiguous, the choice of specific recurrent architecture becomes more critical.

RANKL protein is expressed at the pannus-bone interface at sites of articular bone erosion in rheumatoid arthritis

Objectives. Receptor activator of NF-kappa B ligand (RANKL) and osteoprotegerin (OPG) have been demonstrated to be critical regulators of osteoclast generation and activity. In addition, RANKL has been implicated as an important mediator of bone erosion in rheumatoid arthritis (RA). However, the expression of RANKL and OPG at sites of pannus invasion into bone has not been examined. The present study was undertaken to further elucidate the contribution of this cytokine system to osteoclastogenesis and subsequent bone erosion in RA by examining the pattern of protein expression for RANKL, OPG and the receptor activator of NF-kappa B (RANK) in RA at sites of articular bone erosion. Methods. Tissues from 20 surgical procedures from 17 patients with RA were collected as discarded materials. Six samples contained only synovium or tenosynovium remote from bone, four samples contained pannus-bone interface with adjacent synovium and 10 samples contained both synovium remote from bone and pannu-bone interface with adjacent synovium. Immunohistochemistry was used to characterize the cellular pattern of RANKL, RANK and OPG protein expression immediately adjacent to and remote from sites of bone erosion. Results. Cellular expression of RANKL protein was relatively restricted in the bone microenvironment; staining was focal and confined largely to sites of osteoclast-mediated erosion at the pannus-bone interface and at sites of subchondral bone erosion. RANK-expressing osteoclast precursor cells were also present in these sites. OPG protein expression was observed in numerous cells in synovium remote from bone but was more limited at sites of bone erosion, especially in regions associated with RANKL expression. Conclusions. The pattern of RANKL and OPG expression and the presence of RANK-expressing osteoclast precursor cells at sites of bone erosion in RA contributes to the generation of a local microenvironment that favours osteoclast differentiation and activity. These data provide further evidence implicating RANKL in the pathogenesis of arthritis-induced joint destruction.

Predicting structural disruption of proteins caused by crossover

We present a machine learning model that predicts a structural disruption score from a protein’s primary structure. SCHEMA was introduced by Frances Arnold and colleagues as a method for determining putative recombination sites of a protein on the basis of the full (PDB) description of its structure. The present method provides an alternative to SCHEMA that is able to determine the same score from sequence data only. Circumventing the need for resolving the full structure enables the exploration of yet unresolved and even hypothetical sequences for protein design efforts. Deriving the SCHEMA score from a primary structure is achieved using a two step approach: first predicting a secondary structure from the sequence and then predicting the SCHEMA score from the predicted secondary structure. The correlation coefficient for the prediction is 0.88 and indicates the feasibility of replacing SCHEMA with little loss of precision. ©2005 IEEE

Exploiting sequence dependencies in the prediction of peroxisomal proteins

Prediction of peroxisomal matrix proteins generally depends on the presence of one of two distinct motifs at the end of the amino acid sequence. PTS1 peroxisomal proteins have a well conserved tripeptide at the C-terminal end. However, the preceding residues in the sequence arguably play a crucial role in targeting the protein to the peroxisome. Previous work in applying machine learning to the prediction of peroxisomal matrix proteins has failed W capitalize on the full extent of these dependencies. We benchmark a range of machine learning algorithms, and show that a classifier - based on the Support Vector Machine - produces more accurate results when dependencies between the conserved motif and the preceding section are exploited. We publish an updated and rigorously curated data set that results in increased prediction accuracy of most tested models.

Fuzzy modeling based recognition of multi-font numerals

In this paper, we present a new scheme for off-line recognition of multi-font numerals using the Takagi-Sugeno (TS) model. In this scheme, the binary image of a character is partitioned into a fixed number of sub-images called boxes. The features consist of normalized vector distances (gamma) from each box. Each feature extracted from different fonts gives rise to a fuzzy set. However, when we have a small number of fonts as in the case of multi-font numerals, the choice of a proper fuzzification function is crucial. Hence, we have devised a new fuzzification function involving parameters, which take account of the variations in the fuzzy sets. The new fuzzification function is employed in the TS model for the recognition of multi-font numerals.