UFold: fast and accurate RNA secondary structure prediction with deep learning

For many RNA molecules, the secondary structure is essential for the correct function of the RNA. Predicting RNA secondary structure from nucleotide sequences is a long-standing problem in genomics, but the prediction performance has reached a plateau over time. Traditional RNA secondary structure prediction algorithms are primarily based on thermodynamic models through free energy minimization, which imposes strong prior assumptions and is slow to run. Here, we propose a deep learning-based method, called UFold, for RNA secondary structure prediction, trained directly on annotated data and base-pairing rules. UFold proposes a novel image-like representation of RNA sequences, which can be efficiently processed by Fully Convolutional Networks (FCNs). We benchmark the performance of UFold on both within- and cross-family RNA datasets. It significantly outperforms previous methods on within-family datasets, while achieving a similar performance as the traditional methods when trained and tested on distinct RNA families. UFold is also able to predict pseudoknots accurately. Its prediction is fast with an inference time of about 160 ms per sequence up to 1500 bp in length. An online web server running UFold is available at https://ufold.ics.uci.edu. Code is available at https://github.com/uci-cbcl/UFold.     

Using GEFs to deorphanize odorant receptors

Odorant perception is initiated when odorants activate uniquecombinations of odorant receptors (ORs) expressed in the ciliaof olfactory sensory neurons in the nose (Buck and Axel 1991;Malnic et al. 1999). One of the greatest challenges in the olfactoryscience is to correlate ORs with their ligands (to deorphanizethe ORs). The determination of OR–ligand pairs shouldreveal how the OR family is used to generate diverse odorantperceptions (Keller et al. 2007, Menashi et al. 2007). However,these studies are complicated     

Olfactory receptor antagonism between odorants

The detection of thousands of volatile odorants is mediated by several hundreds of different G protein-coupled olfactory receptors (ORs). The main strategy in encoding odorant identities is a combinatorial receptor code scheme in that different odorants are recognized by different sets of ORs. Despite increasing information on agonist-OR combinations, little is known about the antagonism of ORs in the mammalian olfactory system. Here we show that odorants inhibit odorant responses of OR(s), evidence of antagonism between odorants at the receptor level. The antagonism was demonstrated in a heterologous OR-expression system and in single olfactory neurons that expressed a given OR, and was also visualized at the level of the olfactory epithelium. Dual functions of odorants as an agonist and an antagonist to ORs indicate a new aspect in the receptor code determination for odorant mixtures that often give rise to novel perceptual qualities that are not present in each component. The current study also provides insight into strategies to modulate perceived odorant quality.     

Comparison of odorant specificity of two human olfactory receptors from different phylogenetic classes and evidence for antagonism

Humans are able to detect and discriminate myriads of odorants using only several hundred olfactory receptors (ORs) classified in two major phylogenetic classes representing ORs from aquatic (class I) and terrestrial animals (class II). Olfactory perception results in a combinatorial code, in which one OR recognizes multiple odorants and different odorants are recognized by different combinations of ORs. Moreover, recent data suggest that odorants could also behave as antagonists for other ORs, thus making the combinatorial coding more complex. Here we describe the odorant repertoires of two human ORs belonging to class I and class II, respectively. For this purpose, we set up an assay based on calcium imaging in which 100 odorants were screened using air-phase odorant stimulation at physiological doses. We showed that the human class I OR52D1 is functional, exhibiting a narrow repertoire related to that of its orthologous murine OR, demonstrating than this human class I OR is not an evolutionary relic. The class II OR1G1 was revealed to be broadly tuned towards odorants of 9-10 carbon chain length, with diverse functional groups. The existence of antagonist odorants for the class II OR was also demonstrated. They are structurally related to the agonists, with shorter carbon chain length.     

Combinatorial receptor codes for odors

The discriminatory capacity of the mammalian olfactory system is such that thousands of volatile chemicals are perceived as having distinct odors. Here we used a combination of calcium imaging and single-cell RT-PCR to identify odorant receptors (ORs) for odorants with related structures but varied odors. We found that one OR recognizes multiple odorants and that one odorant is recognized by multiple ORs, but that different odorants are recognized by different combinations of ORs. Thus, the olfactory system uses a combinatorial receptor coding scheme to encode odor identities. Our studies also indicate that slight alterations in an odorant, or a change in its concentration, can change its "code," potentially explaining how such changes can alter perceived odor quality.     

A machine learning approach for the identification of odorant binding proteins from sequence-derived properties

Background  

Odorant binding proteins (OBPs) are believed to shuttle odorants from the environment to the underlying odorant receptors, for which they could potentially serve as odorant presenters. Although several sequence based search methods have been exploited for protein family prediction, less effort has been devoted to the prediction of OBPs from sequence data and this area is more challenging due to poor sequence identity between these proteins.     

PROCAIN: protein profile comparison with assisting information

Detection of remote sequence homology is essential for the accurate inference of protein structure, function and evolution. The most sensitive detection methods involve the comparison of evolutionary patterns reflected in multiple sequence alignments (MSAs) of protein families. We present PROCAIN, a new method for MSA comparison based on the combination of ‘vertical’ MSA context (substitution constraints at individual sequence positions) and ‘horizontal’ context (patterns of residue content at multiple positions). Based on a simple and tractable profile methodology and primitive measures for the similarity of horizontal MSA patterns, the method achieves the quality of homology detection comparable to a more complex advanced method employing hidden Markov models (HMMs) and secondary structure (SS) prediction. Adding SS information further improves PROCAIN performance beyond the capabilities of current state-of-the-art tools. The potential value of the method for structure/function predictions is illustrated by the detection of subtle homology between evolutionary distant yet structurally similar protein domains. ProCAIn, relevant databases and tools can be downloaded from: http://prodata.swmed.edu/procain/download. The web server can be accessed at http://prodata.swmed.edu/procain/procain.php.     

A Digital Atlas of the Dog Brain

There is a long history and a growing interest in the canine as a subject of study in neuroscience research and in translational neurology. In the last few years, anatomical and functional magnetic resonance imaging (MRI) studies of awake and anesthetized dogs have been reported. Such efforts can be enhanced by a population atlas of canine brain anatomy to implement group analyses. Here we present a canine brain atlas derived as the diffeomorphic average of a population of fifteen mesaticephalic dogs. The atlas includes: 1) A brain template derived from in-vivo, T1-weighted imaging at 1 mm isotropic resolution at 3 Tesla (with and without the soft tissues of the head); 2) A co-registered, high-resolution (0.33 mm isotropic) template created from imaging of ex-vivo brains at 7 Tesla; 3) A surface representation of the gray matter/white matter boundary of the high-resolution atlas (including labeling of gyral and sulcal features). The properties of the atlas are considered in relation to historical nomenclature and the evolutionary taxonomy of the Canini tribe. The atlas is available for download (https://cfn.upenn.edu/aguirre/wiki/public:data_plosone_2012_datta).     

A broadly tuned odorant receptor in neurons of trichoid sensilla in locust,Locusta migratoria

Insects have evolved sophisticated olfactory reception systems to sense exogenous chemical signals. Odorant receptors (ORs) on the membrane of chemosensory neurons are believed to be key molecules in sensing exogenous chemical cues. ORs in different species of insects are diverse and should tune a species to its own specific semiochemicals relevant to their survival. The orthopteran insect, locust (Locusta migratoria), is a model hemimetabolous insect. There is very limited knowledge on the functions of locust ORs although many locust OR genes have been identified in genomic sequencing experiments. In this paper, a locust OR, LmigOR3 was localized to neurons housed in trichoid sensilla by in situ hybridization. LmigOR3 was expressed as a transgene in Drosophila trichoid olfactory neurons (aT1) lacking the endogenous receptor Or67d and the olfactory tuning curve and dose-response curves were established for this locust receptor. The results show that LmigOR3 sensitizes neurons to ketones, esters and heterocyclic compounds, indicating that LmigOR3 is a broadly tuned receptor. LmigOR3 is the first odorant receptor from Orthoptera that has been functionally analyzed in the Drosophila aT1 system. This work demonstrates the utility of the Drosophila aT1 system for functional analysis of locust odorant receptors and suggests that LmigOR3 may be involved in detecting food odorants, or perhaps locust body volatiles that may help us to develop new control methods for locusts.     

Odorant receptors in cancer

Odorant receptors (ORs), the largest subfamily of G protein-coupled receptors, detect odorants in the nose. In addition, ORs were recently shown to be expressed in many nonolfactory tissues and cells, indicating that these receptors have physiological and pathophysiological roles beyond olfaction. Many ORs are expressed by tumor cells and tissues, suggesting that they may be associated with cancer progression or may be cancer biomarkers. This review describes OR expression in various types of cancer and the association of these receptors with various types of signaling mechanisms. In addition, the clinical relevance and significance of the levels of OR expression were evaluated. Namely, levels of OR expression in cancer were analyzed based on RNA-sequencing data reported in the Cancer Genome Atlas; OR expression patterns were visualized using t-distributed stochastic neighbor embedding (t-SNE); and the associations between patient survival and levels of OR expression were analyzed. These analyses of the relationships between patient survival and expression patterns obtained from an open mRNA database in cancer patients indicate that ORs may be cancer biomarkers and therapeutic targets.     

SeqNLS: Nuclear Localization Signal Prediction Based on Frequent Pattern Mining and Linear Motif Scoring

Nuclear localization signals (NLSs) are stretches of residues in proteins mediating their importing into the nucleus. NLSs are known to have diverse patterns, of which only a limited number are covered by currently known NLS motifs. Here we propose a sequential pattern mining algorithm SeqNLS to effectively identify potential NLS patterns without being constrained by the limitation of current knowledge of NLSs. The extracted frequent sequential patterns are used to predict NLS candidates which are then filtered by a linear motif-scoring scheme based on predicted sequence disorder and by the relatively local conservation (IRLC) based masking.The experiment results on the newly curated Yeast and Hybrid datasets show that SeqNLS is effective in detecting potential NLSs. The performance comparison between SeqNLS with and without the linear motif scoring shows that linear motif features are highly complementary to sequence features in discerning NLSs. For the two independent datasets, our SeqNLS not only can consistently find over 50% of NLSs with prediction precision of at least 0.7, but also outperforms other state-of-the-art NLS prediction methods in terms of F1 score or prediction precision with similar or higher recall rates. The web server of the SeqNLS algorithm is available at http://mleg.cse.sc.edu/seqNLS.     

Significant sparse polygenic risk scores across 813 traits in UK Biobank

We present a systematic assessment of polygenic risk score (PRS) prediction across more than 1,500 traits using genetic and phenotype data in the UK Biobank. We report 813 sparse PRS models with significant (p < 2.5 x 10⁻⁵) incremental predictive performance when compared against the covariate-only model that considers age, sex, types of genotyping arrays, and the principal component loadings of genotypes. We report a significant correlation between the number of genetic variants selected in the sparse PRS model and the incremental predictive performance (Spearman’s ⍴ = 0.61, p = 2.2 x 10⁻⁵⁹ for quantitative traits, ⍴ = 0.21, p = 9.6 x 10⁻⁴ for binary traits). The sparse PRS model trained on European individuals showed limited transferability when evaluated on non-European individuals in the UK Biobank. We provide the PRS model weights on the Global Biobank Engine (https://biobankengine.stanford.edu/prs).     

Predicted 3-D structures for mouse I7 and rat I7 olfactory receptors and comparison of predicted odor recognition profiles with experiment

The first step in the perception of an odor is the activation of one or more olfactory receptors (ORs) following binding of the odorant molecule to the OR. In order to initiate the process of determining how the molecular level receptor-odorant interactions are related to odor perception, we used the MembStruk computational method to predict the three-dimensional (3-D) structure of the I7 OR for both mouse and rat. We then used the HierDock ligand docking computational method to predict the binding site and binding energy for the library of 56 odorants to these receptors for which experiment response data are now available. We find that the predicted 3-D structures of the mouse and rat I7 OR lead to predictions of odorant binding that are in good agreement with the experimental results, thus validating the accuracy of both the 3-D structure and the predicted binding site. In particular we predict that heptanal and octanal both bind strongly to both mouse and rat I7 ORs, which conflicts with the older literature but agrees with recent experiments. To provide the basis of additional validations of our 3-D structures, we also report the odorant binding site for a new odorant (8-hydroxy-octanal) with a novel functionality designed to bind strongly to mouse I7. Such validated computational methods should be very useful in predicting the structure and function of many other ORs.     

TBI Server: A Web Server for Predicting Ion Effects in RNA Folding

Background

Metal ions play a critical role in the stabilization of RNA structures. Therefore, accurate prediction of the ion effects in RNA folding can have a far-reaching impact on our understanding of RNA structure and function. Multivalent ions, especially Mg²⁺, are essential for RNA tertiary structure formation. These ions can possibly become strongly correlated in the close vicinity of RNA surface. Most of the currently available software packages, which have widespread success in predicting ion effects in biomolecular systems, however, do not explicitly account for the ion correlation effect. Therefore, it is important to develop a software package/web server for the prediction of ion electrostatics in RNA folding by including ion correlation effects.

Results

The TBI web server http://rna.physics.missouri.edu/tbi_index.html provides predictions for the total electrostatic free energy, the different free energy components, and the mean number and the most probable distributions of the bound ions. A novel feature of the TBI server is its ability to account for ion correlation and ion distribution fluctuation effects.

Conclusions

By accounting for the ion correlation and fluctuation effects, the TBI server is a unique online tool for computing ion-mediated electrostatic properties for given RNA structures. The results can provide important data for in-depth analysis for ion effects in RNA folding including the ion-dependence of folding stability, ion uptake in the folding process, and the interplay between the different energetic components.     

Axon guidance of mouse olfactory sensory neurons by odorant receptors and the beta2 adrenergic receptor

Odorant receptors (ORs) provide the core determinant of identity for axons of olfactory sensory neurons (OSNs) to coalesce into glomeruli in the olfactory bulb. Here, using gene targeting in mice, we examine how the OR protein determines axonal identity. An OR::GFP fusion protein is present in axons, consistent with a direct function of ORs in axon guidance. When the OR coding region is deleted, we observe OSNs that coexpress other ORs that function in odorant reception and axonal identity. It remains unclear if such coexpression is normally prevented by negative feedback on OR gene choice. A drastic reduction in OR protein level produces axonal coalescence into novel, remote glomeruli. By contrast, chimeric ORs and ORs with minor mutations perturb axon outgrowth. Strikingly, the beta2 adrenergic receptor can substitute for an OR in glomerular formation when expressed from an OR locus. Thus, ORs have not evolved a unique function in axon guidance.     

Enhancing protein-vitamin binding residues prediction by multiple heterogeneous subspace SVMs ensemble

Background

Vitamins are typical ligands that play critical roles in various metabolic processes. The accurate identification of the vitamin-binding residues solely based on a protein sequence is of significant importance for the functional annotation of proteins, especially in the post-genomic era, when large volumes of protein sequences are accumulating quickly without being functionally annotated.

Results

In this paper, a new predictor called TargetVita is designed and implemented for predicting protein-vitamin binding residues using protein sequences. In TargetVita, features derived from the position-specific scoring matrix (PSSM), predicted protein secondary structure, and vitamin binding propensity are combined to form the original feature space; then, several feature subspaces are selected by performing different feature selection methods. Finally, based on the selected feature subspaces, heterogeneous SVMs are trained and then ensembled for performing prediction.

Conclusions

The experimental results obtained with four separate vitamin-binding benchmark datasets demonstrate that the proposed TargetVita is superior to the state-of-the-art vitamin-specific predictor, and an average improvement of 10% in terms of the Matthews correlation coefficient (MCC) was achieved over independent validation tests. The TargetVita web server and the datasets used are freely available for academic use at http://csbio.njust.edu.cn/bioinf/TargetVita or http://www.csbio.sjtu.edu.cn/bioinf/TargetVita.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2105-15-297) contains supplementary material, which is available to authorized users.     

dbVOR: a database system for importing pedigree,phenotype and genotype data and exporting selected subsets

Background

When studying the genetics of a human trait, we typically have to manage both genome-wide and targeted genotype data. There can be overlap of both people and markers from different genotyping experiments; the overlap can introduce several kinds of problems. Most times the overlapping genotypes are the same, but sometimes they are different. Occasionally, the lab will return genotypes using a different allele labeling scheme (for example 1/2 vs A/C). Sometimes, the genotype for a person/marker index is unreliable or missing. Further, over time some markers are merged and bad samples are re-run under a different sample name. We need a consistent picture of the subset of data we have chosen to work with even though there might possibly be conflicting measurements from multiple data sources.

Results

We have developed the dbVOR database, which is designed to hold data efficiently for both genome-wide and targeted experiments. The data are indexed for fast retrieval by person and marker. In addition, we store pedigree and phenotype data for our subjects. The dbVOR database allows us to select subsets of the data by several different criteria and to merge their results into a coherent and consistent whole. Data may be filtered by: family, person, trait value, markers, chromosomes, and chromosome ranges. The results can be presented in columnar, Mega2, or PLINK format.

Conclusions

dbVOR serves our needs well. It is freely available from https://watson.hgen.pitt.edu/register. Documentation for dbVOR can be found at https://watson.hgen.pitt.edu/register/docs/dbvor.html.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-015-0505-4) contains supplementary material, which is available to authorized users.