Predicting patterns of prey use from morphology of fishes

Synopsis Ecomorphological analyses that search for patterns of association between morphological and prey-use data sets will have a greater chance of understanding the causal relationships between form and diet if the morphological variables used have known consequences for feeding performance. We explore the utility of fish body size, mouth gape and jaw-lever mechanics in predicting patterns of prey use in two very different communities of fishes, Caribbean coral reef fishes, and species of the Centrarchidae that live in Lake Opinicon, Ontario. In spite of major differences in the spectrum of potential prey available, the centrarchids of Lake Opinicon show dietary transitions during ontogeny that are very similar to those seen among and within species of Caribbean groupers (Serranidae). The transition from small zooplankton to intermediate sized invertebrates and ultimately to fishes appears to be very general in ram-suction feeding fishes and is probably driven largely by the constraints of mouth size on prey capture ability. The jaw-lever systems for mouth opening and closing represent direct trade-offs for speed and force of jaw movement. The ratio of in-lever to out-lever in the opening system changes during ontogeny in bluegill, indicating that the mechanics and kinematics of jaw movement may change as well. Among 34 species of Caribbean reef fishes, biting species had jaw-closing ratios that favored force translation, while species that employ rapid-strike ram-suction had closing ratios that enhanced speed of closing and mouth opening ratios that favored a more rapid expansion of the mouth during the strike. We suggest that when prey are categorized into functional groups, reflecting the specific performance features that are important in capturing and handling them, and the differences among habitats in the available prey resource are taken into account, general patterns can be found in morphology-diet relations that cross phylogenetic boundaries.     

The use of evolutionary patterns in protein annotation

With genomic data skyrocketing, their biological interpretation remains a serious challenge. Diverse computational methods address this problem by pointing to the existence of recurrent patterns among sequence, structure, and function. These patterns emerge naturally from evolutionary variation, natural selection, and divergence--the defining features of biological systems--and they identify molecular events and shapes that underlie specificity of function and allosteric communication. Here we review these methods, and the patterns they identify in case studies and in proteome-wide applications, to infer and rationally redesign function.     

Predicting root biomass from branching patterns of Douglas-fir root systems

There are many examples of branching networks in nature, such as tree crowns, river systems, arteries and lungs. These networks have often been described as being self-similar, or following scale-invariant branching rules, and this property has been used to derive several scaling laws. In this paper we model root systems of Douglas-fir ( Pseudotsuga menziesii var. glauca (Beissn.) Franco) as branching networks following several simple branching rules. Our objective is to establish a relationship between trunk diameter and root biomass. We explore the effect of the self-similar branching assumption on this relationship. Using data collected from a mature stand in British Columbia, we find that branching asymmetry and the rate of root taper change with root size, thereby violating the assumption of self-similarity. However, the data are in general agreement with Leonardo da Vinci's area-preserving branching hypothesis. We use the field data to parameterize two models, one assuming self-similar branching and a second incorporating the measured size dependencies of branching parameters. The two models differ by only a small amount (≈8%) in their predictions. For both models, the predicted relationship between trunk diameter and root biomass is in good concordance with previously published empirical data. We conclude that the assumption of self-similar branching, although violated by the data, nevertheless provides a useful tool for predicting the allometric relationship between trunk diameter and root biomass. Finally, we use our models to show that the geometric properties of individual bifurcations fundamentally change the root biomass cost of different root topologies.     

Predicting individuals' learning success from patterns of pre-learning MRI activity

Performance in most complex cognitive and psychomotor tasks improves with training, yet the extent of improvement varies among individuals. Is it possible to forecast the benefit that a person might reap from training? Several behavioral measures have been used to predict individual differences in task improvement, but their predictive power is limited. Here we show that individual differences in patterns of time-averaged T2*-weighted MRI images in the dorsal striatum recorded at the initial stage of training predict subsequent learning success in a complex video game with high accuracy. These predictions explained more than half of the variance in learning success among individuals, suggesting that individual differences in neuroanatomy or persistent physiology predict whether and to what extent people will benefit from training in a complex task. Surprisingly, predictions from white matter were highly accurate, while voxels in the gray matter of the dorsal striatum did not contain any information about future training success. Prediction accuracy was higher in the anterior than the posterior half of the dorsal striatum. The link between trainability and the time-averaged T2*-weighted signal in the dorsal striatum reaffirms the role of this part of the basal ganglia in learning and executive functions, such as task-switching and task coordination processes. The ability to predict who will benefit from training by using neuroimaging data collected in the early training phase may have far-reaching implications for the assessment of candidates for specific training programs as well as the study of populations that show deficiencies in learning new skills.     

Automatic recognition and annotation of gene expression patterns of fly embryos

MOTIVATION: Gene expression patterns obtained by in situ mRNA hybridization provide important information about different genes during Drosophila embryogenesis. So far, annotations of these images are done by manually assigning a subset of anatomy ontology terms to an image. This time-consuming process depends heavily on the consistency of experts. RESULTS: We develop a system to automatically annotate a fruitfly's embryonic tissue in which a gene has expression. We formulate the task as an image pattern recognition problem. For a new fly embryo image, our system answers two questions: (1) Which stage range does an image belong to? (2) Which annotations should be assigned to an image? We propose to identify the wavelet embryo features by multi-resolution 2D wavelet discrete transform, followed by min-redundancy max-relevance feature selection, which yields optimal distinguishing features for an annotation. We then construct a series of parallel bi-class predictors to solve the multi-objective annotation problem since each image may correspond to multiple annotations. SUPPLEMENTARY INFORMATION: The complete annotation prediction results are available at: http://www.cs.niu.edu/~jzhou/papers/fruitfly and http://research.janelia.org/peng/proj/fly_embryo_annotation/. The datasets used in experiments will be available upon request to the correspondence author.     

Automatic recognition and annotation of gene expression patterns of fly embryos

Vol. 23, No.     

It's the machine that matters: Predicting gene function and phenotype from protein networks

Increasing knowledge about the organization of proteins into complexes, systems, and pathways has led to a flowering of theoretical approaches for exploiting this knowledge in order to better learn the functions of proteins and their roles underlying phenotypic traits and diseases. Much of this body of theory has been developed and tested in model organisms, relying on their relative simplicity and genetic and biochemical tractability to accelerate the research. In this review, we discuss several of the major approaches for computationally integrating proteomics and genomics observations into integrated protein networks, then applying guilt-by-association in these networks in order to identify genes underlying traits. Recent trends in this field include a rising appreciation of the modular network organization of proteins underlying traits or mutational phenotypes, and how to exploit such protein modularity using computational approaches related to the internet search algorithm PageRank. Many protein network-based predictions have recently been experimentally confirmed in yeast, worms, plants, and mice, and several successful approaches in model organisms have been directly translated to analyze human disease, with notable recent applications to glioma and breast cancer prognosis.     

Automated annotation of Drosophila gene expression patterns using a controlled vocabulary

MOTIVATION: Regulation of gene expression in space and time directs its localization to a specific subset of cells during development. Systematic determination of the spatiotemporal dynamics of gene expression plays an important role in understanding the regulatory networks driving development. An atlas for the gene expression patterns of fruit fly Drosophila melanogaster has been created by whole-mount in situ hybridization, and it documents the dynamic changes of gene expression pattern during Drosophila embryogenesis. The spatial and temporal patterns of gene expression are integrated by anatomical terms from a controlled vocabulary linking together intermediate tissues developed from one another. Currently, the terms are assigned to patterns manually. However, the number of patterns generated by high-throughput in situ hybridization is rapidly increasing. It is, therefore, tempting to approach this problem by employing computational methods. RESULTS: In this article, we present a novel computational framework for annotating gene expression patterns using a controlled vocabulary. In the currently available high-throughput data, annotation terms are assigned to groups of patterns rather than to individual images. We propose to extract invariant features from images, and construct pyramid match kernels to measure the similarity between sets of patterns. To exploit the complementary information conveyed by different features and incorporate the correlation among patterns sharing common structures, we propose efficient convex formulations to integrate the kernels derived from various features. The proposed framework is evaluated by comparing its annotation with that of human curators, and promising performance in terms of F1 score has been reported.     

Automatic annotation of spatial expression patterns via sparse Bayesian factor models

Advances in reporters for gene expression have made it possible to document and quantify expression patterns in 2D-4D. In contrast to microarrays, which provide data for many genes but averaged and/or at low resolution, images reveal the high spatial dynamics of gene expression. Developing computational methods to compare, annotate, and model gene expression based on images is imperative, considering that available data are rapidly increasing. We have developed a sparse Bayesian factor analysis model in which the observed expression diversity of among a large set of high-dimensional images is modeled by a small number of hidden common factors. We apply this approach on embryonic expression patterns from a Drosophila RNA in situ image database, and show that the automatically inferred factors provide for a meaningful decomposition and represent common co-regulation or biological functions. The low-dimensional set of factor mixing weights is further used as features by a classifier to annotate expression patterns with functional categories. On human-curated annotations, our sparse approach reaches similar or better classification of expression patterns at different developmental stages, when compared to other automatic image annotation methods using thousands of hard-to-interpret features. Our study therefore outlines a general framework for large microscopy data sets, in which both the generative model itself, as well as its application for analysis tasks such as automated annotation, can provide insight into biological questions.     

Predicting CK2 beta-dependent substrates using linear patterns

CK2 is a constitutively active Ser/Thr protein kinase deregulated in cancer and other pathologies, responsible for about the 20% of the human phosphoproteome. The holoenzyme is a complex composed of two catalytic (α or α´) and two regulatory (β) subunits, with individual subunits also coexisting in the cell. In the holoenzyme, CK2β is a substrate-dependent modulator of kinase activity. Therefore, a comprehensive characterization of CK2 cellular function should firstly address which substrates are phosphorylated exclusively when CK2β is present (class-III or beta-dependent substrates). However, current experimental constrains limit this classification to a few substrates. Here, we took advantage of motif-based prediction and designed four linear patterns for predicting class-III behavior in sets of experimentally determined CK2 substrates. Integrating high-throughput substrate prediction, functional classification and network analysis, our results suggest that beta-dependent phosphorylation might exert particular regulatory roles in viral infection and biological processes/pathways like apoptosis, DNA repair and RNA metabolism. It also pointed, that human beta-dependent substrates are mainly nuclear, a few of them shuttling between nuclear and cytoplasmic compartments.     

Functional annotation from predicted protein interaction networks

MOTIVATION: Progress in large-scale experimental determination of protein-protein interaction networks for several organisms has resulted in innovative methods of functional inference based on network connectivity. However, the amount of effort and resources required for the elucidation of experimental protein interaction networks is prohibitive. Previously we, and others, have developed techniques to predict protein interactions for novel genomes using computational methods and data generated from other genomes. RESULTS: We evaluated the performance of a network-based functional annotation method that makes use of our predicted protein interaction networks. We show that this approach performs equally well on experimentally derived and predicted interaction networks, for both manually and computationally assigned annotations. We applied the method to predicted protein interaction networks for over 50 organisms from all domains of life, providing annotations for many previously unannotated proteins and verifying existing low-confidence annotations. AVAILABILITY: Functional predictions for over 50 organisms are available at http://bioverse.compbio.washington.edu and datasets used for analysis at http://data.compbio.washington.edu/misc/downloads/nannotation_data/. SUPPLEMENTARY INFORMATION: A supplemental appendix gives additional details not in the main text. (http://data.compbio.washington.edu/misc/downloads/nannotation_data/supplement.pdf).     

Predicting extinction risk of butterflies and moths (Macrolepidoptera) from distribution patterns and species characteristics

An extensive follow-up study of day- and night-active Macrolepidoptera was performed during 2004 at the Kullaberg Nature Reserve located in the south-western part of Sweden. Butterflies were surveyed in an area of 100 km² and night-active moths were trapped in the core area of the reserve. Macrolepidopteran species resident in the area in the 1950s were compared with species resident in the area in 2004. As much as 159 of 597 species (27%) resident in the area in 1950s were not found at all in 2004 and 22 species (4%) had colonised the area. Butterflies exhibited a disastrous decline with a loss of 45% of the fauna, and day-active species had declined more than night-active species had. Species distribution pattern and species characteristics were used to predict the probability that a species had become extinct or colonised the area. Species limited to one or a few food plants, with a short flight-period or restricted to non-forest habitats were all associated with a high extinction risk. Species occurring in fewer European countries and recorded from few provinces in Sweden were all associated with a higher extinction risk compared with ubiquitous species. For expanding species (colonisers), the best predictor was their distribution area in Sweden and colonising species were more likely to be limited to a few provinces in the southernmost part of Sweden. Species extinct from Kullaberg also have decreased in Finland and species that had colonised Kullaberg also have increased in Finland, while the species with no change in Kullaberg are also relatively constant in Finland. Indeed, the macrolepidopteran fauna was severely reduced in the last 50 years and almost 70% of the habitat specialists were lost. For red-listed species the situation was found to be alarming. Almost 70% of red-listed species, resident in the area in the 1950s, was not found in 2004. This study highlights that species composition changes rapidly even in protected nature reserves and that similar changes in the macrolepidopteran fauna seem to occur over large areas.     

Predicting patterns of post-fire germination in 35 eastern Australian Fabaceae

Germination in 35 species from 15 legume genera of southeastern Australia was promoted by a heat treatment which broke the seed coatcaused dormancy. Once the critical temperature was reached, most seeds had their dormancy broken, independent of the duration of heating. Species fell into three classes according to whether their dormancy was broken by a temperature of 40, 60 or 80°C. Highest germination in all species was achieved by heating in the temperature range 80–100°C, although long durations (120 min) at 100°C caused seed death in several species. At 120°C, seeds of most species were killed at all but one minute's duration. A proportion of seeds from 7 species (Acacia myrtifolia, Pultenaea daphnoides, P. incurvata, P. linophylla, P. polifolia, Dillwynia floribunda and Sphaerolobium vimineurn) was not killed at 120°C and had their dormancy broken. This proportion varied markedly and resultant germination levels were significantly less than those at 80 and 100°C, except in S. vimineum. Between-site variations in the 4 species tested (A. myrtifolia, A. suaveolens, A. terminalis and A. ulicifolia) were small. These variations concerned: (i) the minimum temperature required to break seed dormancy in 2 species: 60°C in one population of A. myrtifolia and A. suaveolens, and 80°C in the other; and (ii) the intensity of the germination response. Duration of heating was less important than temperature as a determinant of germination. Ordination techniques revealed that results from one duration across temperatures were comparable with data from multiple durations. This has significant applications in studying rare species, where seed may be in short supply. Predicted germination levels after a moderate intensity fire should far exceed those after a low intensity fire. Little germination was predicted for many species after a low intensity fire and for one species, A. elongata, no germination was predicted. The potential role of indicator species in relation to the maintenance of species in a community is suggested.     

Predicting patterns of plant species richness in megadiverse South Africa

Using new tools (boosted regression trees) in predictive biogeography, with extensive spatial 23 distribution data for >19 000 species, we developed predictive models for South African plant species richness patterns. Further, biome level analysis explored possible functional determinants of country‐wide regional species richness. Finally, to test model reliability independently, we predicted potential alien invasive plant species richness with an independent dataset. Amongst the different hypotheses generally invoked to explain species 30 diversity (energy, favorableness, topographic heterogeneity, irregularity and seasonality), results revealed topographic heterogeneity as the most powerful single explanatory variable for indigenous South African plant species richness. Some biome‐specific responses were observed, i.e. two of the five analyzed biomes (Fynbos and Grassland) had richness best explained by the “species‐favorableness” hypothesis, but even in this case, topographic heterogeneity was also a primary predictor. This analysis, the largest conducted on an almost exhaustive species sample in a species‐rich region, demonstrates the preeminence of topographic heterogeneity in shaping the spatial pattern of regional plant species richness. Model reliability was confirmed by the considerable predictive power for alien invasive species richness. It thus appears that topographic heterogeneity controls species richness in two main ways: firstly, by providing an abundance of ecological niches in contemporary space (revealed by alien invasive species richness relationships) and secondly, by facilitating the persistence of ecological niches through time. The extraordinary richness of the South African Fynbos biome, a world‐renowned hotspot of biodiversity with the steepest environmental gradients in South Africa, may thus have arisen through both mechanisms. Comparisons with similar regions of the world outside South Africa are needed to confirm the generality of topographic heterogeneity and favorableness as predictors of plant richness.     

Corpus annotation for mining biomedical events from literature

Background  

Advanced Text Mining (TM) such as semantic enrichment of papers, event or relation extraction, and intelligent Question Answering have increasingly attracted attention in the bio-medical domain. For such attempts to succeed, text annotation from the biological point of view is indispensable. However, due to the complexity of the task, semantic annotation has never been tried on a large scale, apart from relatively simple term annotation.     

Adenylate kinase isoenzyme patterns in tissues of people of different phenotype

Three distinct isoenzyme sets have been demonstrated for adenylate kinase (AK) from human tissues by the technique of starch gel electrophoresis. In adult and fetal muscle, liver, kidney, brain, spleen, lung, and leukocytes the same set was found, though individual bands had different intensities. In adult red cells and in fetal red cells separate and distinctive isoenzyme sets were observed. However, in all tissues, adult and fetal, the phenotype of an individual was recognizable by the presence or absence of a characteristic cathodal band. The apparent anomaly of the same phenotype being expressed in different isoenzyme sets was resolved by the demonstration that in adult and fetal red cells AK complexed with hemoglobin. The implications of this complex are briefly discussed.Supported by grants from the Scottish Hospital Endowments Research Trust (HERT 309) and the Muscular Dystrophy Group of Great Britain.     

Automatic annotation of BIND molecular interactions from three-dimensional structures

Software to automate the process of extracting molecular interactions from three-dimensional (3D) structures has been developed that records these as Biomolecular Interaction Network Database (BIND) pairwise interaction records. Full annotation of BIND records is provided through a database processing tool called MMDBind, including detailed atom-atom and residue-residue level interaction information. BIND three-dimensional interaction annotation is synthesized by combining information from the Molecular Modeling Database (MMDB), and the HET (heterogen) group dictionary of small molecules in the macromolecular Crystallographic Information Format (mmCIF). Interactions are validated using the Protein Quaternary Structure (PQS) system. A total of 18,166 interactions were removed as being redundant or biologically irrelevant after PQS validation. This first pass MMDBind annotation creates two new divisions of BIND, 3D Biopolymers (BIND-3DBP) comprising 16,737 initial interaction records, and 3D Small Molecules (BIND-3DSM) comprising 48,219 records. Visualization of interacting residues and nucleotides within a macromolecular structure is possible directly from the BIND database owing to added 3D feature annotation within the BIND records that can be conveniently seen using Cn3D ("see-in-3D") after query from the BIND Data Manager. These interaction records provide a further demonstration of the completeness of the BIND data specification and its capabilities as storage and exchange format for all kinds of molecular interactions, including RNA, DNA, protein, and small molecules. Data from the 3DBP and 3DSM sets are available for downloading in Abstract Syntax Notation.1 (ASN.1) or Extensible Markup Language (XML) formats at ftp://ftp.bind.ca/DB/MMDBBind. Data from the 3DBP set is available for interactive query from the BIND Data Manager at www.bind.ca.