Learning to extract relations for protein annotation

MOTIVATION: Protein annotation is a task that describes protein X in terms of topic Y. Usually, this is constructed using information from the biomedical literature. Until now, most of literature-based protein annotation work has been done manually by human annotators. However, as the number of biomedical papers grows ever more rapidly, manual annotation becomes more difficult, and there is increasing need to automate the process. Recently, information extraction (IE) has been used to address this problem. Typically, IE requires pre-defined relations and hand-crafted IE rules or annotated corpora, and these requirements are difficult to satisfy in real-world scenarios such as in the biomedical domain. In this article, we describe an IE system that requires only sentences labelled according to their relevance or not to a given topic by domain experts. RESULTS: We applied our system to meet the annotation needs of a well-known protein family database; the results show that our IE system can annotate proteins with a set of extracted relations by learning relations and IE rules for disease, function and structure from only relevant and irrelevant sentences.     

Conflicting biomedical assumptions for mathematical modeling: the case of cancer metastasis

Computational models in biomedicine rely on biological and clinical assumptions. The selection of these assumptions contributes substantially to modeling success or failure. Assumptions used by experts at the cutting edge of research, however, are rarely explicitly described in scientific publications. One can directly collect and assess some of these assumptions through interviews and surveys. Here we investigate diversity in expert views about a complex biological phenomenon, the process of cancer metastasis. We harvested individual viewpoints from 28 experts in clinical and molecular aspects of cancer metastasis and summarized them computationally. While experts predominantly agreed on the definition of individual steps involved in metastasis, no two expert scenarios for metastasis were identical. We computed the probability that any two experts would disagree on k or fewer metastatic stages and found that any two randomly selected experts are likely to disagree about several assumptions. Considering the probability that two or more of these experts review an article or a proposal about metastatic cascades, the probability that they will disagree with elements of a proposed model approaches 1. This diversity of conceptions has clear consequences for advance and deadlock in the field. We suggest that strong, incompatible views are common in biomedicine but largely invisible to biomedical experts themselves. We built a formal Markov model of metastasis to encapsulate expert convergence and divergence regarding the entire sequence of metastatic stages. This model revealed stages of greatest disagreement, including the points at which cancer enters and leaves the bloodstream. The model provides a formal probabilistic hypothesis against which researchers can evaluate data on the process of metastasis. This would enable subsequent improvement of the model through Bayesian probabilistic update. Practically, we propose that model assumptions and hunches be harvested systematically and made available for modelers and scientists.     

AGMIAL: implementing an annotation strategy for prokaryote genomes as a distributed system

We have implemented a genome annotation system for prokaryotes called AGMIAL. Our approach embodies a number of key principles. First, expert manual annotators are seen as a critical component of the overall system; user interfaces were cyclically refined to satisfy their needs. Second, the overall process should be orchestrated in terms of a global annotation strategy; this facilitates coordination between a team of annotators and automatic data analysis. Third, the annotation strategy should allow progressive and incremental annotation from a time when only a few draft contigs are available, to when a final finished assembly is produced. The overall architecture employed is modular and extensible, being based on the W3 standard Web services framework. Specialized modules interact with two independent core modules that are used to annotate, respectively, genomic and protein sequences. AGMIAL is currently being used by several INRA laboratories to analyze genomes of bacteria relevant to the food-processing industry, and is distributed under an open source license.     

Improving the Caenorhabditis elegans genome annotation using machine learning

For modern biology, precise genome annotations are of prime importance, as they allow the accurate definition of genic regions. We employ state-of-the-art machine learning methods to assay and improve the accuracy of the genome annotation of the nematode Caenorhabditis elegans. The proposed machine learning system is trained to recognize exons and introns on the unspliced mRNA, utilizing recent advances in support vector machines and label sequence learning. In 87% (coding and untranslated regions) and 95% (coding regions only) of all genes tested in several out-of-sample evaluations, our method correctly identified all exons and introns. Notably, only 37% and 50%, respectively, of the presently unconfirmed genes in the C. elegans genome annotation agree with our predictions, thus we hypothesize that a sizable fraction of those genes are not correctly annotated. A retrospective evaluation of the Wormbase WS120 annotation [] of C. elegans reveals that splice form predictions on unconfirmed genes in WS120 are inaccurate in about 18% of the considered cases, while our predictions deviate from the truth only in 10%-13%. We experimentally analyzed 20 controversial genes on which our system and the annotation disagree, confirming the superiority of our predictions. While our method correctly predicted 75% of those cases, the standard annotation was never completely correct. The accuracy of our system is further corroborated by a comparison with two other recently proposed systems that can be used for splice form prediction: SNAP and ExonHunter. We conclude that the genome annotation of C. elegans and other organisms can be greatly enhanced using modern machine learning technology.     

Automated Analysis and Reannotation of Subcellular Locations in Confocal Images from the Human Protein Atlas

The Human Protein Atlas contains immunofluorescence images showing subcellular locations for thousands of proteins. These are currently annotated by visual inspection. In this paper, we describe automated approaches to analyze the images and their use to improve annotation. We began by training classifiers to recognize the annotated patterns. By ranking proteins according to the confidence of the classifier, we generated a list of proteins that were strong candidates for reexamination. In parallel, we applied hierarchical clustering to group proteins and identified proteins whose annotations were inconsistent with the remainder of the proteins in their cluster. These proteins were reexamined by the original annotators, and a significant fraction had their annotations changed. The results demonstrate that automated approaches can provide an important complement to visual annotation.     

The institute for genomic research Osa1 rice genome annotation database

We have developed a rice (Oryza sativa) genome annotation database (Osa1) that provides structural and functional annotation for this emerging model species. Using the sequence of O. sativa subsp. japonica cv Nipponbare from the International Rice Genome Sequencing Project, pseudomolecules, or virtual contigs, of the 12 rice chromosomes were constructed. Our most recent release, version 3, represents our third build of the pseudomolecules and is composed of 98% finished sequence. Genes were identified using a series of computational methods developed for Arabidopsis (Arabidopsis thaliana) that were modified for use with the rice genome. In release 3 of our annotation, we identified 57,915 genes, of which 14,196 are related to transposable elements. Of these 43,719 non-transposable element-related genes, 18,545 (42.4%) were annotated with a putative function, 5,777 (13.2%) were annotated as encoding an expressed protein with no known function, and the remaining 19,397 (44.4%) were annotated as encoding a hypothetical protein. Multiple splice forms (5,873) were detected for 2,538 genes, resulting in a total of 61,250 gene models in the rice genome. We incorporated experimental evidence into 18,252 gene models to improve the quality of the structural annotation. A series of functional data types has been annotated for the rice genome that includes alignment with genetic markers, assignment of gene ontologies, identification of flanking sequence tags, alignment with homologs from related species, and syntenic mapping with other cereal species. All structural and functional annotation data are available through interactive search and display windows as well as through download of flat files. To integrate the data with other genome projects, the annotation data are available through a Distributed Annotation System and a Genome Browser. All data can be obtained through the project Web pages at http://rice.tigr.org.     

A database for G proteins and their interaction with GPCRs

Background

The explosion in biological information creates the need for databases that are easy to develop, easy to maintain and can be easily manipulated by annotators who are most likely to be biologists. However, deployment of scalable and extensible databases is not an easy task and generally requires substantial expertise in database development.

Results

BioBuilder is a Zope-based software tool that was developed to facilitate intuitive creation of protein databases. Protein data can be entered and annotated through web forms along with the flexibility to add customized annotation features to protein entries. A built-in review system permits a global team of scientists to coordinate their annotation efforts. We have already used BioBuilder to develop Human Protein Reference Database http://www.hprd.org, a comprehensive annotated repository of the human proteome. The data can be exported in the extensible markup language (XML) format, which is rapidly becoming as the standard format for data exchange.

Conclusions

As the proteomic data for several organisms begins to accumulate, BioBuilder will prove to be an invaluable platform for functional annotation and development of customizable protein centric databases. BioBuilder is open source and is available under the terms of LGPL.     

Small membrane proteins found by comparative genomics and ribosome binding site models

The correct annotation of genes encoding the smallest proteins is one of the biggest challenges of genome annotation, and perhaps more importantly, few annotated short open reading frames have been confirmed to correspond to synthesized proteins. We used sequence conservation and ribosome binding site models to predict genes encoding small proteins, defined as having 16–50 amino acids, in the intergenic regions of the Escherichia coli genome. We tested expression of these predicted as well as previously annotated genes by integrating the sequential peptide affinity tag directly upstream of the stop codon on the chromosome and assaying for synthesis using immunoblot assays. This approach confirmed that 20 previously annotated and 18 newly discovered proteins of 16–50 amino acids are synthesized. We summarize the properties of these small proteins; remarkably more than half of the proteins are predicted to be single‐transmembrane proteins, nine of which we show co‐fractionate with cell membranes.     

Annotated Chemical Patent Corpus: A Gold Standard for Text Mining

Exploring the chemical and biological space covered by patent applications is crucial in early-stage medicinal chemistry activities. Patent analysis can provide understanding of compound prior art, novelty checking, validation of biological assays, and identification of new starting points for chemical exploration. Extracting chemical and biological entities from patents through manual extraction by expert curators can take substantial amount of time and resources. Text mining methods can help to ease this process. To validate the performance of such methods, a manually annotated patent corpus is essential. In this study we have produced a large gold standard chemical patent corpus. We developed annotation guidelines and selected 200 full patents from the World Intellectual Property Organization, United States Patent and Trademark Office, and European Patent Office. The patents were pre-annotated automatically and made available to four independent annotator groups each consisting of two to ten annotators. The annotators marked chemicals in different subclasses, diseases, targets, and modes of action. Spelling mistakes and spurious line break due to optical character recognition errors were also annotated. A subset of 47 patents was annotated by at least three annotator groups, from which harmonized annotations and inter-annotator agreement scores were derived. One group annotated the full set. The patent corpus includes 400,125 annotations for the full set and 36,537 annotations for the harmonized set. All patents and annotated entities are publicly available at www.biosemantics.org.     

Annotation of selection strengths in viral genomes

MOTIVATION: Viral genomes tend to code in overlapping reading frames to maximize informational content. This may result in atypical codon bias and particular evolutionary constraints. Due to the fast mutation rate of viruses, there is additional strong evidence for varying selection between intra- and intergenomic regions. The presence of multiple coding regions complicates the concept of K(a)/K(s) ratio, and thus begs for an alternative approach when investigating selection strengths. Building on the paper by McCauley and Hein, we develop a method for annotating a viral genome coding in overlapping reading frames. We introduce an evolutionary model capable of accounting for varying levels of selection along the genome, and incorporate it into our prior single sequence HMM methodology, extending it now to a phylogenetic HMM. Given an alignment of several homologous viruses to a reference sequence, we may thus achieve an annotation both of coding regions as well as selection strengths, allowing us to investigate different selection patterns and hypotheses. RESULTS: We illustrate our method by applying it to a multiple alignment of four HIV2 sequences, as well as of three Hepatitis B sequences. We obtain an annotation of the coding regions, as well as a posterior probability for each site of the strength of selection acting on it. From this we may deduce the average posterior selection acting on the different genes. Whilst we are encouraged to see in HIV2, that the known to be conserved genes gag and pol are indeed annotated as such, we also discover several sites of less stringent negative selection within the env gene. To the best of our knowledge, we are the first to subsequently provide a full selection annotation of the Hepatitis B genome by explicitly modelling the evolution within overlapping reading frames, and not relying on simple K(a)/K(s) ratios.     

Heterogeneity in multistage carcinogenesis and mixture modeling

Carcinogenesis is commonly described as a multistage process, in which stem cells are transformed into cancer cells via a series of mutations. In this article, we consider extensions of the multistage carcinogenesis model by mixture modeling. This approach allows us to describe population heterogeneity in a biologically meaningful way. We focus on finite mixture models, for which we prove identifiability. These models are applied to human lung cancer data from several birth cohorts. Maximum likelihood estimation does not perform well in this application due to the heavy censoring in our data. We thus use analytic graduation instead. Very good fits are achieved for models that combine a small high risk group with a large group that is quasi immune.     

On analyzing circadian rhythms data using nonlinear mixed models with harmonic terms

Wang, Ke, and Brown (2003, Biometrics59, 804-812) developed a smoothing-based approach for modeling circadian rhythms with random effects. Their approach is flexible in that fixed and random covariates can affect both the amplitude and phase shift of a nonparametrically smoothed periodic function. In motivating their approach, Wang et al. stated that a simple sinusoidal function is too restrictive. In addition, they stated that "although adding harmonics can improve the fit, it is difficult to decide how many harmonics to include in the model, and the results are difficult to interpret." We disagree with the notion that harmonic models cannot be a useful tool in modeling longitudinal circadian rhythm data. In this note, we show how nonlinear mixed models with harmonic terms allow for a simple and flexible alternative to Wang et al.'s approach. We show how to choose the number of harmonics using penalized likelihood to flexibly model circadian rhythms and to estimate the effect of covariates on the rhythms. We fit harmonic models to the cortisol circadian rhythm data presented by Wang et al. to illustrate our approach. Furthermore, we evaluate the properties of our procedure with a small simulation study. The proposed parametric approach provides an alternative to Wang et al.'s semiparametric approach and has the added advantage of being easy to implement in most statistical software packages.     

Modeling Inter‐Subject Variability in fMRI Activation Location: A Bayesian Hierarchical Spatial Model

Summary The aim of this article is to develop a spatial model for multi‐subject fMRI data. There has been extensive work on univariate modeling of each voxel for single and multi‐subject data, some work on spatial modeling of single‐subject data, and some recent work on spatial modeling of multi‐subject data. However, there has been no work on spatial models that explicitly account for inter‐subject variability in activation locations. In this article, we use the idea of activation centers and model the inter‐subject variability in activation locations directly. Our model is specified in a Bayesian hierarchical framework which allows us to draw inferences at all levels: the population level, the individual level, and the voxel level. We use Gaussian mixtures for the probability that an individual has a particular activation. This helps answer an important question that is not addressed by any of the previous methods: What proportion of subjects had a significant activity in a given region. Our approach incorporates the unknown number of mixture components into the model as a parameter whose posterior distribution is estimated by reversible jump Markov chain Monte Carlo. We demonstrate our method with a fMRI study of resolving proactive interference and show dramatically better precision of localization with our method relative to the standard mass‐univariate method. Although we are motivated by fMRI data, this model could easily be modified to handle other types of imaging data.     

Mass spectrometry-based prokaryote gene annotation

MS combined with database searching has become the preferred method for identifying proteins present in cell or tissue samples. The technique enables us to execute large-scale proteome analyses of species whose genomes have already been sequenced. Searching mass spectrometric data against protein databases composed of annotated genes has been widely conducted. However, there are some issues with this technique; wrong annotations in protein databases cause deterioration in the accuracy of protein identification, and only proteins that have already been annotated can be identified. We propose a new framework that can detect correct ORFs by integrating an MS/MS proteomic data mapping and a knowledge-based system regarding the translation initiation sites. This technique can provide correction of predicted coding sequences, together with the possibility of identifying novel genes. We have developed a computational system; it should first conduct the probabilistic peptide-matching against all possible translational frames using MS/MS data, then search for discriminative DNA patterns around the detected peptides, and lastly integrate the facts using empirical knowledge stored in knowledge bases to obtain correct ORFs. We used photosynthetic bacteria Synechocystis sp. PCC6803 as a sample prokaryote, resulting in the finding of 14 N-terminus annotation errors and several new candidate genes.     

A semantic analysis of the annotations of the human genome

The correct interpretation of any biological experiment depends in an essential way on the accuracy and consistency of the existing annotation databases. Such databases are ubiquitous and used by all life scientists in most experiments. However, it is well known that such databases are incomplete and many annotations may also be incorrect. In this paper we describe a technique that can be used to analyze the semantic content of such annotation databases. Our approach is able to extract implicit semantic relationships between genes and functions. This ability allows us to discover novel functions for known genes. This approach is able to identify missing and inaccurate annotations in existing annotation databases, and thus help improve their accuracy. We used our technique to analyze the current annotations of the human genome. From this body of annotations, we were able to predict 212 additional gene-function assignments. A subsequent literature search found that 138 of these gene-functions assignments are supported by existing peer-reviewed papers. An additional 23 assignments have been confirmed in the meantime by the addition of the respective annotations in later releases of the Gene Ontology database. Overall, the 161 confirmed assignments represent 75.95% of the proposed gene-function assignments. Only one of our predictions (0.4%) was contradicted by the existing literature. We could not find any relevant articles for 50 of our predictions (23.58%). The method is independent of the organism and can be used to analyze and improve the quality of the data of any public or private annotation database.     

A large‐scale proteogenomics study of apicomplexan pathogens—Toxoplasma gondii and Neospora caninum

Proteomics data can supplement genome annotation efforts, for example being used to confirm gene models or correct gene annotation errors. Here, we present a large‐scale proteogenomics study of two important apicomplexan pathogens: Toxoplasma gondii and Neospora caninum. We queried proteomics data against a panel of official and alternate gene models generated directly from RNASeq data, using several newly generated and some previously published MS datasets for this meta‐analysis. We identified a total of 201 996 and 39 953 peptide‐spectrum matches for T. gondii and N. caninum, respectively, at a 1% peptide FDR threshold. This equated to the identification of 30 494 distinct peptide sequences and 2921 proteins (matches to official gene models) for T. gondii, and 8911 peptides/1273 proteins for N. caninum following stringent protein‐level thresholding. We have also identified 289 and 140 loci for T. gondii and N. caninum, respectively, which mapped to RNA‐Seq‐derived gene models used in our analysis and apparently absent from the official annotation (release 10 from EuPathDB) of these species. We present several examples in our study where the RNA‐Seq evidence can help in correction of the current gene model and can help in discovery of potential new genes. The findings of this study have been integrated into the EuPathDB. The data have been deposited to the ProteomeXchange with identifiers PXD000297and PXD000298.