A hybrid SOM-SVM approach for the zebrafish gene expression analysis

Microarray technology can be employed to quantitatively measure the expression of thousands of genes in a single experiment. It has become one of the main tools for global gene expression analysis in molecular biology research in recent years. The large amount of expression data generated by this technology makes the study of certain complex biological problems possible, and machine learning methods are expected to play a crucial role in the analysis process. In this paper, we present our results from integrating the self-organizing map （SOM） and the support vector machine （SVM） for the analysis of the various functions of zebrafish genes based on their expression. The most distinctive characteristic of our zebrafish gene expression is that the number of samples of different classes is imbalanced. We discuss how SOM can be used as a data-filtering tool to improve the classification performance of the SVM on this data set.     

A new hybrid approach to predict subcellular localization of proteins by incorporating gene ontology

Based on the recent development in the gene ontology and functional domain databases, a new hybridization approach is developed for predicting protein subcellular location by combining the gene product, functional domain, and quasi-sequence-order effects. As a showcase, the same prokaryotic and eukaryotic datasets, which were studied by many previous investigators, are used for demonstration. The overall success rate by the jackknife test for the prokaryotic set is 94.7% and that for the eukaryotic set 92.9%. These are so far the highest success rates achieved for the two datasets by following a rigorous cross-validation test procedure, suggesting that such a hybrid approach may become a very useful high-throughput tool in the area of bioinformatics, proteomics, as well as molecular cell biology. The very high success rates also reflect the fact that the subcellular localization of a protein is closely correlated with: (1). the biological objective to which the gene or gene product contributes, (2). the biochemical activity of a gene product, and (3). the place in the cell where a gene product is active.     

A hybrid sequence approach to the paracelsus challenge

Inspired by the Paracelsus Challenge of Rose and Creamer (Proteins 19:1–3, 1994), we have designed a protein sequence that is 50% identical to an all-helical protein but is intended to fold into a largely β-sheet structure. Rather than attempt a de novo design, our strategy was to construct a hybrid sequence based on a helical “parent” protein (434 Cro) and a “target” protein with the desired fold (the B1 domain of protein G). The hybrid sequence (Crotein-G) is 50% identical to 434 Cro but is also 62% identical to the B1 domain of protein G. We also created a variant of Crotein-G (ZCrotein-G) that contains a potential His₃Cys₁ zinc binding site. At low protein concentrations and in the presence of 20% 2,2,2-trifluoroethanol (TFE) (v/v), the circular dichroism spectra of the designed proteins are distinct from that of 434 Cro and similar to that of the B1 domain of protein G. However, the proteins fail to denature in a cooperative manner. Furthermore, aggregation occurs at moderate protein concentrations or in the absence of TFE. Addition of zinc to ZCrotein-G does not promote structure formation. In summary, 434 Cro has been altered to something that may resemble the B1 domain of protein G, but the protein does not adopt a native structure. Proteins 30:136–143, 1998. © 1998 Wiley-Liss, Inc.     

A genomic approach to identify hybrid incompatibility genes

Uncovering the genetic and molecular basis of barriers to gene flow between populations is key to understanding how new species are born. Intrinsic postzygotic reproductive barriers such as hybrid sterility and hybrid inviability are caused by deleterious genetic interactions known as hybrid incompatibilities. The difficulty in identifying these hybrid incompatibility genes remains a rate-limiting step in our understanding of the molecular basis of speciation. We recently described how whole genome sequencing can be applied to identify hybrid incompatibility genes, even from genetically terminal hybrids. Using this approach, we discovered a new hybrid incompatibility gene, gfzf, between Drosophila melanogaster and Drosophila simulans, and found that it plays an essential role in cell cycle regulation. Here, we discuss the history of the hunt for incompatibility genes between these species, discuss the molecular roles of gfzf in cell cycle regulation, and explore how intragenomic conflict drives the evolution of fundamental cellular mechanisms that lead to the developmental arrest of hybrids.     

A rational approach to nasal reconstruction in leprosy

Destruction of the nasal septum and nasal bones by Mycobacterium leprae and subsequent infection is still seen regularly in leprosy endemic areas. The social stigma associated with this deformity is significant. Many different procedures have been developed to reconstruct the nose. Patients operated on at Anandaban Hospital and the Green Pastures Hospital and Rehabilitation Center between 1986 and 2001 were reviewed. There were 48 patients with an average age of 47 years. Five deformities were mild, 22 were moderate, 13 were severe, and eight were not graded. Bone grafting with nasolabial skin flaps was performed in 14 cases, bone grafting alone was performed in 10 cases, flaps alone were performed in seven cases, and cartilage grafting was performed in 10 cases. In three patients, a prosthesis was inserted, and in three patients a gull-wing forehead flap was performed. Overall, excellent or good cosmetic results were obtained in 83 percent of cases. Grafting with conchal cartilage was associated with the best cosmetic results and had minimal complications. Bone grafting with or without nasolabial flaps was associated with a 50 percent complication rate of infection or graft resorption. In mild to moderate deformities, cartilage grafting is recommended; for more severe deformities, bone grafting with bony fixation and skin flaps is recommended. Perioperative antibiotics must be used, and these procedures should be performed by an experienced surgeon. In very severe cases with skin deficiency, reconstruction with a forehead flap gives good results.     

A hybrid approach to theoretical analysis of bovine pancreatic trypsin inhibitor.

A static analysis of bovine pancreatic trypsin inhibitor (BPTI) is presented based on a new discrete/continuum approach to modeling the dynamics of biomolecules. This hybrid method utilizes knowledge of the intramolecular potential and molecular configuration to generate a field of elastic modulus tensors. These tensors, which relate the local stress and strain for each atom in the biomolecule, can be used to judge the local rigidity as well as indicate regions of high stress. Comparing the tensor fields for an unrelaxed and a relaxed configuration, the microscopic structure of BPTI is found to be anisotropic and to have regions of stress even when it is relaxed in the potential field. However, when these fields are averaged over the whole protein or over individual residues the structure becomes more isotropic and the stressed regions vanish. Using these averaged tensors, we calculated bulk properties such as Young's modulus and the Lamé constants and they agreed with previously reported values.     

The acquired auricular deformity. A systematic approach to its analysis and reconstruction.

Acquired ear deformities present unique and varied problems which tax the ingenuity of the plastic surgeon. Case individualization is necessary, and a systematic assessment of the residual tissues is a requisite when planning an appropriate reconstruction. Cases performed by one surgeon have been used to demonstrate the basic principles of these repairs. A systematic approach to the reconstruction of these acquired deformities is presented.     

QuickJoin--fast neighbour-joining tree reconstruction

We have built a tool for fast construction of very large phylogenetic trees. The tool uses heuristics for speeding up the neighbour-joining algorithm-while still constructing the same tree as the original neighbour-joining algorithm-making it possible to construct trees for 8000 species in <10 min on a single desktop PC. In comparison, the same task takes more than 30 min using the QuickTree neighbour-joining implementation.     

A coalescence approach to gene conversion

In this paper we develop a coalescent model with intralocus gene conversion. Such models are of increasing importance in the analysis of intralocus variability and linkage disequilibrium. We derive the distribution of the waiting time until a gene conversion event occurs in a sample in terms of the distribution of the length of the transferred segment, zeta. We do not assume any specific form of the distribution of zeta. Further, given that a gene conversion event occurs we find the distribution of (sigma, tau), the end points of the transferred segment and derive results on correlations between local trees in positions chi(1) and chi(2). Among other results we show that the correlation between the branch lengths of two local trees in the coalescent with gene conversion (and no recombination) decreases toward a nonzero constant when the distance between chi(1) and chi(2) increases. Finally, we show that a model including both recombination and gene conversion might account for the lack of intralocus associations found in, e.g., Drosophila melanogaster.     

The hybrid resolution approach to automated autoradiographic analysis.

The use of radioactive isotopes in the investigation of cells and their relations to an organism as a physiologic unit is an important area of research to the biologist. The manual analysis of the autoradiographs which result from the use of radioactive tracers is, however, very time consuming and prone to error. This paper describes the utilization of hybrid resolution image processing techniques for the rapid and automatic analysis of tissue region autoradiographs. Nine regional features can be extracted and six properties characterized using the present system implementation. Priliminary measurements on two tissue types show good agreement between manual and automated readings.     

Distance-based reconstruction of tree models for oncogenesis.

Comparative genomic hybridization (CGH) is a laboratory method to measure gains and losses in the copy number of chromosomal regions in tumor cells. It is hypothesized that certain DNA gains and losses are related to cancer progression and that the patterns of these changes are relevant to the clinical consequences of the cancer. It is therefore of interest to develop models which predict the occurrence of these events, as well as techniques for learning such models from CGH data. We continue our study of the mathematical foundations for inferring a model of tumor progression from a CGH data set that we started in Desper et al. (1999). In that paper, we proposed a class of probabilistic tree models and showed that an algorithm based on maximum-weight branching in a graph correctly infers the topology of the tree, under plausible assumptions. In this paper, we extend that work in the direction of the so-called distance-based trees, in which events are leaves of the tree, in the style of models common in phylogenetics. Then we show how to reconstruct the distance-based trees using tree-fitting algorithms developed by researchers in phylogenetics. The main advantages of the distance-based models are that 1) they represent information about co-occurrences of all pairs of events, instead of just some pairs, 2) they allow quantitative predictions about which events occur early in tumor progression, and 3) they bring into play the extensive methodology and software developed in the context of phylogenetics. We illustrate the distance-based tree method and how it complements the branching tree method, with a CGH data set for renal cancer.     

On the repeat-annotated phylogenetic tree reconstruction problem.

A new problem in phylogenetic inference is presented, based on recent biological findings indicating a strong association between reversals (i.e., inversions) and repeats. These biological findings are formalized here in a new mathematical model, called repeat-annotated phylogenetic trees (RAPT). We show that, under RAPT, the evolutionary process--including both the tree-topology as well as internal node genome orders--is uniquely determined, a property that is of major significance both in theory and in practice. Furthermore, the repeats are employed to provide linear-time algorithms for reconstructing both the genomic orders and the phylogeny, which are NP-hard problems under the classical model of sorting by reversals (SBR).     

The average common substring approach to phylogenomic reconstruction.

We describe a novel method for efficient reconstruction of phylogenetic trees, based on sequences of whole genomes or proteomes, whose lengths may greatly vary. The core of our method is a new measure of pairwise distances between sequences. This measure is based on computing the average lengths of maximum common substrings, which is intrinsically related to information theoretic tools (Kullback-Leibler relative entropy). We present an algorithm for efficiently computing these distances. In principle, the distance of two l long sequences can be calculated in O(l) time. We implemented the algorithm using suffix arrays our implementation is fast enough to enable the construction of the proteome phylogenomic tree for hundreds of species and the genome phylogenomic forest for almost two thousand viruses. An initial analysis of the results exhibits a remarkable agreement with "acceptable phylogenetic and taxonomic truth." To assess our approach, our results were compared to the traditional (single-gene or protein-based) maximum likelihood method. The obtained trees were compared to implementations of a number of alternative approaches, including two that were previously published in the literature, and to the published results of a third approach. Comparing their outcome and running time to ours, using a "traditional" trees and a standard tree comparison method, our algorithm improved upon the "competition" by a substantial margin. The simplicity and speed of our method allows for a whole genome analysis with the greatest scope attempted so far. We describe here five different applications of the method, which not only show the validity of the method, but also suggest a number of novel phylogenetic insights.     

Comparing two Bayesian methods for gene tree/species tree reconstruction: simulations with incomplete lineage sorting and horizontal gene transfer

With the increasing interest in recognizing the discordance between gene genealogies, various gene tree/species tree reconciliation methods have been developed. We present here the first attempt to assess and compare two such Bayesian methods, Bayesian estimation of species trees (BEST) and BUCKy (Bayesian untangling of concordance knots), in the presence of several known processes of gene tree discordance. DNA alignments were simulated under the influence of incomplete lineage sorting (ILS) and of horizontal gene transfer (HGT). BEST and BUCKy both account for uncertainty in gene tree estimation but differ substantially in their assumptions of what caused gene tree discordance. BEST estimates a species tree using the coalescent model, assuming that all gene tree discordance is due to ILS. BUCKy does not assume any specific biological process of gene tree discordance through the use of a nonparametric clustering of concordant genes. BUCKy estimates the concordance factor (CF) of a clade, which is defined as the proportion of genes that truly have the clade in their trees. The estimated concordance tree is then built from clades with the highest estimated CFs. Because of their different assumptions, it was expected that BEST would perform better in the presence of ILS and that BUCKy would perform better in the presence of HGT. As expected, the species tree was more accurately reconstructed by BUCKy in the presence of HGT, when the HGT events were unevenly placed across the species tree. BUCKy and BEST performed similarly in most other cases, including in the presence of strong ILS and of HGT events that were evenly placed across the tree. However, BUCKy was shown to underestimate the uncertainty in CF estimation, with short credibility intervals. Despite this, the discordance pattern estimated by BUCKy could be compared with the signature of ILS. The resulting test for the adequacy of the coalescent model proved to have low Type I error. It was powerful when HGT was the major source of discordance and when HGT events were unevenly placed across the species tree.