TOP MEDICAL JOURNALS

Background: The combination of gene expression profiling with linkage analysis has become a powerful paradigm for mapping gene expression quantitative trait loci (eQTL). To date, most studies have searched for eQTL by analyzing gene expression traits one at a time. As thousands of expression traits are typically analyzed, this can reduce power because of the need to correct for the number of hypothesis tests performed. In addition, gene expression traits exhibit a complex correlation structure, which is unutilized when analyzing traits individually. Results: To address these issues, we applied two different multivariate dimension reduction techniques, the Singular Value Decomposition (SVD) and Independent Component Analysis (ICA) to gene expression traits derived from a cross between two strains of Saccharomyces cerevisiae. Both methods decompose the data into a set of meta-traits, which are linear combinations of all the expression traits. The meta-traits were enriched for several Gene Ontology categories including metabolic pathways, stress response, RNA processing, ion transport, retro-transposition and telomeric maintenance. Genome-wide linkage analysis was performed on the top 20 meta-traits from both techniques. In total, 21 eQTL were found, of which 11 are novel. Interestingly, both cis and trans -linkages to the meta-traits were observed. Conclusions: These results demonstrate that dimension reduction methods are a useful and complementary approach for probing the genetic architecture of gene expression variation.

Background: Publicly available DNA sequence databases such as GenBank are large, and are growing at an exponential rate. The sheer volume of data being dealt with presents serious storage and data communications problems. Currently, sequence data is usually kept in large “flat files,” which are then compressed using standard Lempel-Ziv (gzip) compression - an approach which rarely achieves good compression ratios. While much research has been done on compressing individual DNA sequences, surprisingly little has focused on the compression of entire databases of such sequences. In this study we introduce the sequence database compression software COIL. Results: We have designed and implemented a portable software package, COIL, for compressing and decompressing DNA sequence databases based on the idea of edit-tree coding. COIL is geared towards achieving high compression ratios at the expense of execution time and memory usage during compression - the compression time represents a “one-off investment” whose cost is quickly amortised if the resulting compressed file is transmitted many times. Decompression requires little memory and is extremely fast. We demonstrate a 5% improvement in compression ratio over state-of-the-art general-purpose compression tools for a large GenBank database file containing Expressed Sequence Tag (EST) data. Finally, COIL can efficiently encode incremental additions to a sequence database. Conclusions: COIL presents a compelling alternative to conventional compression of flat files for the storage and distribution of DNA sequence databases having a narrow distribution of sequence lengths, such as EST data. Increasing compression levels for databases having a wide distribution of sequence lengths is a direction for future work.

Background: Bioactivity profiling through high-throughput in vitro assays can reduce the cost and time required in toxicological screening of environmental chemicals and can also reduce the need for animal testing. Several public efforts are aimed at discovering patterns or classifiers in high-dimensional bioactivity space that predict tissue, organ or whole animal toxicological endpoints. Supervised machine learning is a powerful approach to discover combinatorial relationships in complex in vitro / in vivo datasets. We present a novel model to simulate complex chemical-toxicology data sets and use this model to evaluate the relative performance of different machine learning (ML) methods. Results: The classification performance of Artificial Neural Networks (ANN), K-Nearest Neighbors (KNN), Linear Discriminant Analysis (LDA), Naive Bayes (NB), Recursive Partitioning and Regression Trees (RPART), and Support Vector Machines (SVM) in the presence and absence of filter-based feature selection was analyzed using K-way cross-validation testing and independent validation on simulated in vitro assay data sets with varying levels of model complexity, number of irrelevant features and measurement noise. While the prediction accuracy of all ML methods decreased as non-causal (irrelevant) features were added, some ML methods performed better than others. In the limit of using a large number of features, ANN and SVM were always in the top performing set of methods while RPART and KNN (k=5) were always in the poorest performing set. The addition of measurement noise irrelevant features decreased the classification accuracy of all ML methods, with LDA suffering the greatest performance degradation. LDA performance is especially sensitive to the use of feature selection. Filter-based feature selection generally improved performance, most strikingly for LDA. Conclusions: We have developed a novel simulation model to evaluate machine learning methods for the analysis of data sets in which in vitro bioassay data is being used to predict in vivo chemical toxicology. From our analysis, we can recommend that several ML methods, most notably SVM and ANN are good candidates for use in real world applications in this area.

Background: The current surgical therapy of midfacial fractures involves internal fixation in which bone fragments are fixed in their anatomical positions with osteosynthesis plates and corresponding screws until bone healing is complete. This often causes new fractures to fragile bones while drilling pilot holes or trying to insert screws. The adhesive fixation of osteosynthesis plates using PMMA bone cement could offer a viable alternative for fixing the plates without screws. In order to achieve the adhesive bonding of bone cement to cortical bone in the viscerocranium, an amphiphilic bone bonding agent was created, analogous to the dentin bonding agents currently on the market. Methods: The adhesive bonding strengths were measured using tension tests. For this, metal plates with 2.0 mm diameter screw holes were cemented with PMMA bone cement to cortical bovine bone samples from the femur diaphysis. The bone was conditioned with an amphiphilic bone bonding agent prior to cementing. The samples were stored for 1 to 42 days at 37 degrees C, either moist or completely submerged in an isotonic NaCl-solution, and then subjected to the tension tests. Results: Without the bone bonding agent, the bonding strength was close to zero (0.2 MPa). Primary stability with bone bonding agent is considered to be at ca. 8 MPa. Moist storage over 42 days resulted in decreased adhesion forces of ca. 6 MPa. Wet storage resulted in relatively constant bonding strengths of ca. 8 MPa. Conclusions: A new amphiphilic bone bonding agent was developed, which builds an optimizied interlayer between the hydrophilic bone surface and the hydrophobic PMMA bone cement and thus leads to adhesive bonding between them. Our in vitro investigations demonstrated the adhesive bonding of PMMA bone cement to cortical bone, which was also stable against hydrolysis. The newly developed adhesive fixing technique could be applied clinically when the fixation of osteosynthesis plates with screws is impossible. With the detected adhesion forces of ca. 6 to 8 MPa, it is assumed that the adhesive fixation system is able to secure bone fragments from the non-load bearing midfacial regions in their orthotopic positions until fracture consolidation is complete.