Identification of protein functions using a machine-learning approach based on sequence-derived properties

Background  

Predicting the function of an unknown protein is an essential goal in bioinformatics. Sequence similarity-based approaches are widely used for function prediction; however, they are often inadequate in the absence of similar sequences or when the sequence similarity among known protein sequences is statistically weak. This study aimed to develop an accurate prediction method for identifying protein function, irrespective of sequence and structural similarities.     

A Statistical Model of Protein Sequence Similarity and Function Similarity Reveals Overly-Specific Function Predictions

Background

Predicting protein function from primary sequence is an important open problem in modern biology. Not only are there many thousands of proteins of unknown function, current approaches for predicting function must be improved upon. One problem in particular is overly-specific function predictions which we address here with a new statistical model of the relationship between protein sequence similarity and protein function similarity.

Methodology

Our statistical model is based on sets of proteins with experimentally validated functions and numeric measures of function specificity and function similarity derived from the Gene Ontology. The model predicts the similarity of function between two proteins given their amino acid sequence similarity measured by statistics from the BLAST sequence alignment algorithm. A novel aspect of our model is that it predicts the degree of function similarity shared between two proteins over a continuous range of sequence similarity, facilitating prediction of function with an appropriate level of specificity.

Significance

Our model shows nearly exact function similarity for proteins with high sequence similarity (bit score >244.7, e-value >1e⁻⁶², non-redundant NCBI protein database (NRDB)) and only small likelihood of specific function match for proteins with low sequence similarity (bit score <54.6, e-value <1e⁻⁰⁵, NRDB). For sequence similarity ranges in between our annotation model shows an increasing relationship between function similarity and sequence similarity, but with considerable variability. We applied the model to a large set of proteins of unknown function, and predicted functions for thousands of these proteins ranging from general to very specific. We also applied the model to a data set of proteins with previously assigned, specific functions that were electronically based. We show that, on average, these prior function predictions are more specific (quite possibly overly-specific) compared to predictions from our model that is based on proteins with experimentally determined function.     

On the Green’s function for a one-dimensional random walk

The Green’s function of a random walk on a lattice is defined as the inverse of the operatorK-z1, whereK is the matrix of transition rates and z is an arbitrary complex parameter. The Green’s function for a symmetrical random walk in one dimension is here explicitly given in closed form for reflecting, periodic, and absorbing boundaries, and also for an infinite lattice.     

Molecular Basis for Oligomeric-DNA Binding and Episome Maintenance by KSHV LANA

LANA is the KSHV-encoded terminal repeat binding protein essential for viral replication and episome maintenance during latency. We have determined the X-ray crystal structure of LANA C-terminal DNA binding domain (LANA_DBD) to reveal its capacity to form a decameric ring with an exterior DNA binding surface. The dimeric core is structurally similar to EBV EBNA1 with an N-terminal arm that regulates DNA binding and is required for replication function. The oligomeric interface between LANA dimers is dispensable for single site DNA binding, but is required for cooperative DNA binding, replication function, and episome maintenance. We also identify a basic patch opposite of the DNA binding surface that is responsible for the interaction with BRD proteins and contributes to episome maintenance function. The structural features of LANA_DBD suggest a novel mechanism of episome maintenance through DNA-binding induced oligomeric assembly.     

Ordering process of self-organizing maps improved by asymmetric neighborhood function

The Self-organizing map (SOM) is an unsupervised learning method based on the neural computation, which has found wide applications. However, the learning process sometime takes multi-stable states, within which the map is trapped to an undesirable disordered state including topological defects on the map. These topological defects critically aggravate the performance of the SOM. In order to overcome this problem, we propose to introduce an asymmetric neighborhood function for the SOM algorithm. Compared with the conventional symmetric one, the asymmetric neighborhood function accelerates the ordering process even in the presence of the defect. However, this asymmetry tends to generate a distorted map. This can be suppressed by an improved method of the asymmetric neighborhood function. In the case of one-dimensional SOM, it is found that the required steps for perfect ordering is numerically shown to be reduced from O(N ³) to O(N ²). We also discuss the ordering process of a twisted state in two-dimensional SOM, which can not be rectified by the ordinary symmetric neighborhood function.     

Four notions of biological function

I argue that there are at least four different ways in which the term ‘function’ is used in connection with the study of living organisms, namely: (1) function as (mere) activity, (2) function as biological role, (3) function as biological advantage, and (4) function as selected effect. Notion (1) refers to what an item does by itself; (2) refers to the contribution of an item or activity to a complex activity or capacity of an organism; (3) refers to the value for the organism of an item having a certain character rather than another; (4) refers to the way in which a trait acquired and has maintained its current share in the population. The recognition of a separate notion of function as biological advantage solves the problem of the indeterminate reference situation that has been raised against a counterfactual analysis of function, and emphasizes the importance of counterfactual comparison in the explanatory practice of organismal biology. This reveals a neglected problem in the philosophy of biology, namely that of accounting for the insights provided by counterfactual comparison.     

Foot-thumping as an alarm signal in macropodoid marsupials: prevalence and hypotheses of function

• 1 Alarm signalling as a means to reduce predation risk is an important component of the behavioural repertoire of many species. It has previously been noted that many of the macropodoid marsupials (kangaroos, wallabies and rat‐kangaroos) produce a foot‐thump, an audible signal created by striking the ground with one or both feet, that is most likely an alarm signal.
• 2 The prevalence of foot‐thumping within the macropodoids and hypotheses of its function as an alarm signal have been poorly documented. To address this issue, we investigate the prevalence of foot‐thumping in macropodoids and interpret possible function according to current alarm signalling theory. Evidence for foot‐thumping was found in almost all macropodoids. In light of this, the behaviour appears to be a conservative trait that may have arisen alongside or followed the evolution of bipedal locomotion, and suggests that this trait carries significant benefits that transcend ecological and predation differences among species.
• 3 Nine alarm signal hypotheses were explored in order to determine the function of foot‐thumping in macropodoid marsupials. However, the existing evidence for consistent function remains inconclusive. Therefore, a series of predictions were developed to provide the foundation for future research to investigate more thoroughly the function of foot‐thumping in macropodoid marsupials.

     

The Role of Cooperation in the Development of Intelligence

In the theoretical analysis of the general structure of activity as presented in the fundamental work of L. S. Vygotsky and A. N. Leont'ev, cooperation among the participants in a joint activity is viewed as the source of new mental functions. Under conditions of cooperation, an activity that is initially shared by those participating in it emerges as an original and fundamental foundation for the development of forms of individual activity. Thus, L. S. Vygotsky wrote, "Any function in the cultural development of the child appears twice, on two levels: first as a social function, and then as a psychological function, first as something taking place between people, a category based on relationship between minds, an intermental category; later, this psychological function occurs within the child, as an intramental category" [2. Pp. 197-98].     

Learning,AMPA receptor mobility and synaptic plasticity depend on n‐cofilin‐mediated actin dynamics

Neuronal plasticity is an important process for learning, memory and complex behaviour. Rapid remodelling of the actin cytoskeleton in the postsynaptic compartment is thought to have an important function for synaptic plasticity. However, the actin‐binding proteins involved and the molecular mechanisms that in vivo link actin dynamics to postsynaptic physiology are not well understood. Here, we show that the actin filament depolymerizing protein n‐cofilin is controlling dendritic spine morphology and postsynaptic parameters such as late long‐term potentiation and long‐term depression. Loss of n‐cofilin‐mediated synaptic actin dynamics in the forebrain specifically leads to impairment of all types of associative learning, whereas exploratory learning is not affected. We provide evidence for a novel function of n‐cofilin function in synaptic plasticity and in the control of extrasynaptic excitatory AMPA receptors diffusion. These results suggest a critical function of actin dynamics in associative learning and postsynaptic receptor availability.     

ESTIMATING THE FORM OF NATURAL SELECTION ON A QUANTITATIVE TRAIT

The fitness function f relates fitness of individuals to the quantitative trait under natural selection. The function is useful in predicting fitness differences among individuals and in revealing whether an optimum is present within the range of phenotypes in the population. It may also be thought of as describing the ecological environment in terms of the trait. Quadratic regression will approximate the fitness function from data (e.g., Lande and Arnold, 1983), but the method does not reliably indicate features of f such as the presence of modes (stabilizing selection) or dips (disruptive selection). I employ an alternative procedure requiring no a priori model for the function. The method is useful in two ways: it provides a nonparametric estimate of f, of interest by itself, and it can be used to suggest an appropriate parametric model. I also discuss measures of selection intensity based on the fitness function. Analysis of six data sets yields a variety of forms of f and provides new insights for some familiar cases. Low amounts of variation and a low density of data points near the tails of many phenotype distributions emerge as limitations to gaining information on fitness functions. An experimental approach in which the distribution of a quantitative trait is broadened through manipulation would minimize these problems.     

Age is an independent risk factor for left atrial dysfunction: results from an observational study

Introduction. The degenerative changes of myocardial tissue are thought to influence left atrial (LA) function. Changes of left atrial function are generally due to changes in left ventricle (LV) compliance. But valvular dysfunction and hypertension as comorbidity cannot be ignored. Women have a different clinical profile compared with men concerning the risk of heart failure. We investigated the influence of increasing age and gender corrected for comorbidity, on left atrial function. Methods. Using an open access echocardiography database, supplemented with additional LA function measurements, we defined three different LA function parameters. Odds ratios (OR) were calculated to reproduce the relation between age, gender and LA function. The association between age, gender and LA function was estimated, and corrected for comorbid conditions as valve disease, high blood pressure and LV dysfunction, using logistic regression. Results. Higher age was positively correlated with increased LA volume, decreased ejection fraction and increased LA kinetic energy. Age per decade increase, corrected for comorbidity, resulted in an increased risk of LA dysfunction (OR between 1.5 and 1.9). Gender had little influence on LA function parameters except for LA maximal volume. Men had a significantly larger LA maximal volume compared with women. Conclusions. In this open access echocardiography database, increasing age was correlated with LA dysfunction. Age per decade increase, corrected for comorbid conditions such as mitral and aortic valve disease, hypertension and heart failure, is an independent risk factor for LA dysfunction. The gender influence on LA dysfunction seems to be limited. (Neth Heart J 2010;18:243-7.)     

An atlas of posttranslational modifications on RNA binding proteins

RNA structure and function are intimately tied to RNA binding protein recognition and regulation. Posttranslational modifications are chemical modifications which can control protein biology. The role of PTMs in the regulation RBPs is not well understood, in part due to a lacking analysis of PTM deposition on RBPs. Herein, we present an analysis of posttranslational modifications (PTMs) on RNA binding proteins (RBPs; a PTM RBP Atlas). We curate published datasets and primary literature to understand the landscape of PTMs and use protein–protein interaction data to understand and potentially provide a framework for understanding which enzymes are controlling PTM deposition and removal on the RBP landscape. Intersection of our data with The Cancer Genome Atlas also provides researchers understanding of mutations that would alter PTM deposition. Additional characterization of the RNA–protein interface provided from in-cell UV crosslinking experiments provides a framework for hypotheses about which PTMs could be regulating RNA binding and thus RBP function. Finally, we provide an online database for our data that is easy to use for the community. It is our hope our efforts will provide researchers will an invaluable tool to test the function of PTMs controlling RBP function and thus RNA biology.     

Effects of ziram on tumor-cell-binding capacity,cell-surface marker expression,and ATP levels of human natural killer cells

Human natural killer (NK) cells are central in immune defense against tumor and virally infected cells. Ziram is used as an accelerating agent in latex production and as an agricultural fungicide. Previous studies showed that continuous exposure to ziram inhibits NK lytic function. Additionally, they showed that a brief (1 h) exposure to ziram caused persistent loss of lytic function. This study examined whether decreases in lytic function were accompanied by decreases in the target-binding function of NK cells and found that some, but not all, exposures to ziram caused significant decreases in binding function. Ziram exposures that caused a loss of binding function were examined for effects on expression of key NK cell-surface proteins needed for binding to targets. Exposure to 2 μM ziram for 1 h followed by 24 or 48 h in ziram-free media decreased CD16 expression, but no other exposures caused decreases in cell-surface proteins. As decreases in adenosine triphosphate (ATP) could be in part responsible for loss of lytic function, the effect of ziram exposures on ATP levels of NK cells were examined. Certain ziram exposures decreased ATP levels in NK cells, but a decrease in ATP was not necessarily associated with a decrease in lytic function. The results indicate that ziram-induced losses of lytic function cannot be fully explained by alteration in binding, cell-surface protein expression, or ATP levels     

Analysis of putative miRNA function using a novel approach, GAPPS-miRTarGE

Deciphering the function of miRNA is one of the most important research subjects directed toward understanding the regulation of gene expression. Several experimental methodologies and bioinformatics programs have been developed, however, elucidating miRNA function has not been an easy task. Herein, we suggest a new method, GAPPS-miRTar^GE, which is a novel methodology for predicting miRNA function based on the proportion of mRNA targets expressed during embryonic developmental stages, the Theilers stages (TS), in mice. GAPPS-miRTar^GE is essentially a computational approach that groups miRNAs using shared expression patterns of their target genes during the 28 different TS. In this study, we present not only several examples derived from the GAPPS-miRTar^GE analyses that confirm previously known miRNA functions but also examples of function prediction for valid but functionally unknown miRNAs. Furthermore, we show that tissue-centered GAPPS-miRTar^GE, such as brain-centered or heart-centered, is useful for predicting miRNA function on a more detailed level.     

Crystal structure of the Vibrio cholerae colonization factor TcpF and identification of a functional immunogenic site

Vibrio cholerae relies on two main virulence factors—toxin-coregulated pilus (TCP) and cholera toxin—to cause the gastrointestinal disease cholera. TCP is a type IV pilus that mediates bacterial autoagglutination and colonization of the intestine. TCP is encoded by the tcp operon, which also encodes TcpF, a protein of unknown function that is secreted by V. cholerae in a TCP-dependent manner. Although TcpF is not required for TCP biogenesis, a tcpF mutant has a colonization defect in the infant mouse cholera model that is as severe as a pilus mutant. Furthermore, TcpF antisera protect against V. cholerae infection. TcpF has no apparent sequence homology to any known protein. Here, we report the de novo X-ray crystal structure of TcpF and the identification of an epitope that is critical for its function as a colonization factor. A monoclonal antibody recognizing this epitope is protective against V. cholerae challenge and adds to the protection provided by an anti-TcpA antibody. These data suggest that TcpF has a novel function in V. cholerae colonization and define a region crucial for this function.     

The Effect of Prenatal and Childhood Development on Hearing,Vision and Cognition in Adulthood

It is unclear what the contribution of prenatal versus childhood development is for adult cognitive and sensory function and age-related decline in function. We examined hearing, vision and cognitive function in adulthood according to self-reported birth weight (an index of prenatal development) and adult height (an index of early childhood development). Subsets (N = 37,505 to 433,390) of the UK Biobank resource were analysed according to visual and hearing acuity, reaction time and fluid IQ. Sensory and cognitive performance was reassessed after ~4 years (N = 2,438 to 17,659). In statistical modelling including age, sex, socioeconomic status, educational level, smoking, maternal smoking and comorbid disease, adult height was positively associated with sensory and cognitive function (partial correlations; pr 0.05 to 0.12, p < 0.001). Within the normal range of birth weight (10^th to 90^th percentile), there was a positive association between birth weight and sensory and cognitive function (pr 0.06 to 0.14, p < 0.001). Neither adult height nor birth weight was associated with change in sensory or cognitive function. These results suggest that adverse prenatal and childhood experiences are a risk for poorer sensory and cognitive function and earlier development of sensory and cognitive impairment in adulthood. This finding could have significant implications for preventing sensory and cognitive impairment in older age.     

Improving the specificity of high-throughput ortholog prediction

Background  

Orthologs (genes that have diverged after a speciation event) tend to have similar function, and so their prediction has become an important component of comparative genomics and genome annotation. The gold standard phylogenetic analysis approach of comparing available organismal phylogeny to gene phylogeny is not easily automated for genome-wide analysis; therefore, ortholog prediction for large genome-scale datasets is typically performed using a reciprocal-best-BLAST-hits (RBH) approach. One problem with RBH is that it will incorrectly predict a paralog as an ortholog when incomplete genome sequences or gene loss is involved. In addition, there is an increasing interest in identifying orthologs most likely to have retained similar function.