A feature selection method for classification within functional genomics experiments based on the proportional overlapping score

Background

Microarray technology, as well as other functional genomics experiments, allow simultaneous measurements of thousands of genes within each sample. Both the prediction accuracy and interpretability of a classifier could be enhanced by performing the classification based only on selected discriminative genes. We propose a statistical method for selecting genes based on overlapping analysis of expression data across classes. This method results in a novel measure, called proportional overlapping score (POS), of a feature’s relevance to a classification task.

Results

We apply POS, along‐with four widely used gene selection methods, to several benchmark gene expression datasets. The experimental results of classification error rates computed using the Random Forest, k Nearest Neighbor and Support Vector Machine classifiers show that POS achieves a better performance.

Conclusions

A novel gene selection method, POS, is proposed. POS analyzes the expressions overlap across classes taking into account the proportions of overlapping samples. It robustly defines a mask for each gene that allows it to minimize the effect of expression outliers. The constructed masks along‐with a novel gene score are exploited to produce the selected subset of genes.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2105-15-274) contains supplementary material, which is available to authorized users.     

Discovery and structural characterization of an allosteric inhibitor of bacterial cis‐prenyltransferase

Undecaprenyl pyrophosphate synthase (UPPs) is an essential enzyme in a key bacterial cell wall synthesis pathway. It catalyzes the consecutive condensations of isopentenyl pyrophosphate (IPP) groups on to a trans-farnesyl pyrophosphate (FPP) to produce a C55 isoprenoid, undecaprenyl pyrophosphate (UPP). Here we report the discovery and co-crystal structures of a drug-like UPPs inhibitor in complex with Streptococcus pneumoniae UPPs, with and without substrate FPP, at resolutions of 2.2 and 2.1 Å, respectively. The UPPs inhibitor has a low molecular weight (355 Da), but displays potent inhibition of UPP synthesis in vitro (IC₅₀ 50 nM) that translates into excellent whole cell antimicrobial activity against pathogenic strains of Streptococcal species (MIC₉₀ 0.4 µg mL⁻¹). Interestingly, the inhibitor does not compete with the substrates but rather binds at a site adjacent to the FPP binding site and interacts with the tail of the substrate. Based on the structures, an allosteric inhibition mechanism of UPPs is proposed for this inhibitor. This inhibition mechanism is supported by biochemical and biophysical experiments, and provides a basis for the development of novel antibiotics targeting Streptococcus pneumoniae.     

Semicontinuous penicillin production by two Penicillium chrysogenum strains immobilized in calcium alginate beads

Summary The growth rates of immobilized Penicillium chrysogenum strains are important in their application to semicontinuous penicillin production. Immobilized P. chrysogenum strains produced about 10–15% less biomass but about 1–2 times more penicillin than free suspended mycelia.In a chemically defined medium an industrial P. chrysogenum strain, S₁, produced about 10–12 times more penicillin than strain ATCC 12690. In a complex medium the immobilized P. chrysogenum S₁ produced about 12% penicillin more than in shaken cultures. In bubble column fermentations, penicillin production was 163% higher in the complex medium than in the chemically defined medium.     

Development of an expressed sequence tag (EST) resource for wheat (Triticum aestivum L.): EST generation, unigene analysis, probe selection and bioinformatics for a 16,000-locus bin-delineated map

This report describes the rationale, approaches, organization, and resource development leading to a large-scale deletion bin map of the hexaploid (2n = 6x = 42) wheat genome (Triticum aestivum L.). Accompanying reports in this issue detail results from chromosome bin-mapping of expressed sequence tags (ESTs) representing genes onto the seven homoeologous chromosome groups and a global analysis of the entire mapped wheat EST data set. Among the resources developed were the first extensive public wheat EST collection (113,220 ESTs). Described are protocols for sequencing, sequence processing, EST nomenclature, and the assembly of ESTs into contigs. These contigs plus singletons (unassembled ESTs) were used for selection of distinct sequence motif unigenes. Selected ESTs were rearrayed, validated by 5′ and 3′ sequencing, and amplified for probing a series of wheat aneuploid and deletion stocks. Images and data for all Southern hybridizations were deposited in databases and were used by the coordinators for each of the seven homoeologous chromosome groups to validate the mapping results. Results from this project have established the foundation for future developments in wheat genomics.     

Quinazolinone-based rhodanine-3-acetic acids as potent aldose reductase inhibitors: Synthesis,functional evaluation and molecular modeling study

A series of quinazolinone-based rhodanine-3-acetic acids was synthesized and tested for in vitro aldose reductase inhibitory activity. All the target compounds displayed nanomolar activity against the target enzyme. Compounds 3a, 3b, and 3e exhibited almost 3-fold higher activity as compared to the only marketed reference drug epalrestat. Structure-activity relationship studies indicated that bulky substituents at the 3-phenyl ring of the quinazolinone moiety are generally not tolerated in the active site of the enzyme. Insertion of a methoxy group on the central benzylidene ring was found to have a variable effect on ALR-2 activity depending on the nature of peripheral quinazolinone ring substituents. Removal of the acetic acid moiety led to inactive or weakly active target compounds. Docking and molecular dynamic simulations of the most active rhodanine-3-acetic acid derivatives were also carried out, to provide the basis for further structure-guided design of novel inhibitors.     

Immunoliposomes bearing lymphocyte activation gene 3 fusion protein and P5 peptide: A novel vaccine for breast cancer

LAG3-Ig as an immune adjuvant has elicited potent anti-tumor immune responses in several preclinical and clinical studies, but the full potential immunostimulatory of LAG3-Ig has yet to be achieved. We hypothesized that by anchoring LAG3-Ig to the surface of liposomes, the adjuvant activity of LAG3-Ig could be improved. We also investigated the immunotherapy by co-delivery of liposome-coupled LAG3-Ig and P5 tumor antigen in mice model of TUBO breast cancer. We prepared and characterized novel PEGylated liposomes bearing surface conjugated LAG3-Ig and P5. Consistent with our hypothesis, liposomes-conjugated LAG3-Ig via multivalent binding to MHC class II molecules exerted immunostimulatory of LAG3-Ig and markedly induced maturation of dendritic cells more efficiently than free LAG3-Ig. LAG3-Ig-P5-immunoliposomes effectively elicited protective anti-tumor responses more than locally injected soluble LAG3-Ig + P5. The higher percentage of CD4⁺ and CD8⁺ T cells in the spleen and more rapid and pronounced infiltration of these effector cells into the site of the tumor were seen following immunoliposome therapy. Finally, anti-tumor immunity induced by LAG3-Ig-P5-immunoliposomes translated into the more tumor regression and prolonged survival of treated mice, compared to soluble immunotherapy. Taken together, our findings suggest that LAG3-Ig-P5-immunoliposomes can be considered as a valuable candidate for developing a liposome-based therapeutic cancer vaccine in treating HER2/ neu⁺ breast cancer patients.     

Antifreeze proteins in overwintering plants: a tale of two activities

Antifreeze proteins are found in a wide range of overwintering plants where they inhibit the growth and recrystallization of ice that forms in intercellular spaces. Unlike antifreeze proteins found in fish and insects, plant antifreeze proteins have multiple, hydrophilic ice-binding domains. Surprisingly, antifreeze proteins from plants are homologous to pathogenesis-related proteins and also provide protection against psychrophilic pathogens. In winter rye (Secale cereale), antifreeze proteins accumulate in response to cold, short daylength, dehydration and ethylene, but not pathogens. Transferring single genes encoding antifreeze proteins to freezing-sensitive plants lowered their freezing temperatures by approximately 1 degrees C. Genes encoding dual-function plant antifreeze proteins are excellent models for use in evolutionary studies to determine how genes acquire new expression patterns and how proteins acquire new activities.     

Activation of the pyrin inflammasome by intracellular Burkholderia cenocepacia

Burkholderia cenocepacia is an opportunistic pathogen that causes chronic infection and induces progressive respiratory inflammation in cystic fibrosis patients. Recognition of bacteria by mononuclear cells generally results in the activation of caspase-1 and processing of IL-1β, a major proinflammatory cytokine. In this study, we report that human pyrin is required to detect intracellular B. cenocepacia leading to IL-1β processing and release. This inflammatory response involves the host adapter molecule ASC and the bacterial type VI secretion system (T6SS). Human monocytes and THP-1 cells stably expressing either small interfering RNA against pyrin or YFP-pyrin and ASC (YFP-ASC) were infected with B. cenocepacia and analyzed for inflammasome activation. B. cenocepacia efficiently activates the inflammasome and IL-1β release in monocytes and THP-1. Suppression of pyrin levels in monocytes and THP-1 cells reduced caspase-1 activation and IL-1β release in response to B. cenocepacia challenge. In contrast, overexpression of pyrin or ASC induced a robust IL-1β response to B. cenocepacia, which correlated with enhanced host cell death. Inflammasome activation was significantly reduced in cells infected with T6SS-defective mutants of B. cenocepacia, suggesting that the inflammatory reaction is likely induced by an as yet uncharacterized effector(s) of the T6SS. Together, we show for the first time, to our knowledge, that in human mononuclear cells infected with B. cenocepacia, pyrin associates with caspase-1 and ASC forming an inflammasome that upregulates mononuclear cell IL-1β processing and release.