Amino acid classification based spectrum kernel fusion for protein subnuclear localization

Intelligent Protein Design and Molecular Characterization Techniques: A Comprehensive Review

In recent years, the widespread application of artificial intelligence algorithms in protein structure, function prediction, and de novo protein design has significantly accelerated the process of intelligent protein design and led to many noteworthy achievements. This advancement in protein intelligent design holds great potential to accelerate the development of new drugs, enhance the efficiency of biocatalysts, and even create entirely new biomaterials. Protein characterization is the key to the performance of intelligent protein design. However, there is no consensus on the most suitable characterization method for intelligent protein design tasks. This review describes the methods, characteristics, and representative applications of traditional descriptors, sequence-based and structure-based protein characterization. It discusses their advantages, disadvantages, and scope of application. It is hoped that this could help researchers to better understand the limitations and application scenarios of these methods, and provide valuable references for choosing appropriate protein characterization techniques for related research in the field, so as to better carry out protein research.

Evaluation of genetic diversity and management of disease in Border Collie dogs

Maintaining genetic diversity in dog breeds is an important consideration for the management of inherited diseases. We evaluated genetic diversity in Border Collies using molecular and genealogical methods, and examined changes to genetic diversity when carriers for Trapped Neutrophil Syndrome (TNS) and Neuronal Ceroid Lipofuscinosis (NCL) are removed from the genotyped population. Genotype data for 255 Border Collies and a pedigree database of 83,996 Border Collies were used for analysis. Molecular estimates revealed a mean multi-locus heterozygosity (MLH) of 0.311 (SD 0.027), 20.79% of the genome consisted of runs of homozygosity (ROH ) > 1 Mb, effective population size (N_e) was 84.7, and mean inbreeding (F) was 0.052 (SD 0.083). For 227 genotyped Border Collies that had available pedigree information (GenoPed), molecular and pedigree estimates of diversity were compared. A reference population (dogs born between 2005 and 2015, inclusive; N = 13,523; RefPop) and their ancestors (N = 12,478) were used to evaluate the diversity of the population that are contributing to the current generation. The reference population had a N_e of 123.5, a mean F of 0.095 (SD 0.082), 2276 founders (f), 205.5 effective founders (f_e), 28 effective ancestors (f_a) and 10.65 (SD 2.82) founder genomes (N_g). Removing TNS and NCL carriers from the genotyped population had a small impact on diversity measures (ROH > 1 Mb, MLH, heterozygosity), however, there was a loss of > 10% minor allele frequency for 89 SNPs around the TNS mutation (maximum loss of 12.7%), and a loss of > 5% for 5 SNPs around the NCL mutation (maximum 5.18%). A common ancestor was identified for 38 TNS-affected dogs and 64 TNS carriers, and a different common ancestor was identified for 33 NCL-affected dogs and 28 carriers, with some overlap of prominent individuals between both pedigrees. Overall, Border Collies have a high level of genetic diversity compared to other breeds.

Multi-Omics Data Fusion via a Joint Kernel Learning Model for Cancer Subtype Discovery and Essential Gene Identification

The multiple sources of cancer determine its multiple causes, and the same cancer can be composed of many different subtypes. Identification of cancer subtypes is a key part of personalized cancer treatment and provides an important reference for clinical diagnosis and treatment. Some studies have shown that there are significant differences in the genetic and epigenetic profiles among different cancer subtypes during carcinogenesis and development. In this study, we first collect seven cancer datasets from the Broad Institute GDAC Firehose, including gene expression profile, isoform expression profile, DNA methylation expression data, and survival information correspondingly. Furthermore, we employ kernel principal component analysis (PCA) to extract features for each expression profile, convert them into three similarity kernel matrices by Gaussian kernel function, and then fuse these matrices as a global kernel matrix. Finally, we apply it to spectral clustering algorithm to get the clustering results of different cancer subtypes. In the experimental results, besides using the P-value from the Cox regression model and survival analysis as the primary evaluation measures, we also introduce statistical indicators such as Rand index (RI) and adjusted RI (ARI) to verify the performance of clustering. Then combining with gene expression profile, we obtain the differential expression of genes among different subtypes by gene set enrichment analysis. For lung cancer, GMPS, EPHA10, C10orf54, and MAGEA6 are highly expressed in different subtypes; for liver cancer, CMYA5, DEPDC6, FAU, VPS24, RCBTB2, LOC100133469, and SLC35B4 are significantly expressed in different subtypes.

ML-rRBF-ECOC: A Multi-Label Learning Classifier for Predicting Protein Subcellular Localization with Both Single and Multiple Sites

Background:

The subcellular localization of a protein is closely related with its functions and interactions. More and more evidences show that proteins may simultaneously exist at, or move between, two or more different subcellular localizations. Therefore, predicting protein subcellular localization is an important but challenging problem.

Observation:

Most of the existing methods for predicting protein subcellular localization assume that a protein locates at a single site. Although a few methods have been proposed to deal with proteins with multiple sites, correlations between subcellular localization are not efficiently taken into account. In this paper, we propose an integrated method for predicting protein subcellular localizations with both single site and multiple sites.

Methods:

Firstly, we extend the Multi-Label Radial Basis Function (ML-RBF) method to the regularized version, and augment the first layer of ML-RBF to take local correlations between subcellular localization into account. Secondly, we embed the modified ML-RBF into a multi-label Error-Correcting Output Codes (ECOC) method in order to further consider the subcellular localization dependency. We name our method ML-rRBF-ECOC. Finally, the performance of ML-rRBF-ECOC is evaluated on three benchmark datasets.

Results:

The results demonstrate that ML-rRBF-ECOC has highly competitive performance to the related multi-label learning method and some state-of-the-art methods for predicting protein subcellular localizations with multiple sites. Considering dependency between subcellular localizations can contribute to the improvement of prediction performance.

Conclusion:

This also indicates that correlations between different subcellular localizations really exist. Our method at least plays a complementary role to existing methods for predicting protein subcellular localizations with multiple sites.

Prediction of Protein Sub-Mitochondria Locations Using Protein Interaction Networks

Background

Prediction of the protein localization is among the most important issues in the bioinformatics that is used for the prediction of the proteins in the cells and organelles such as mitochondria. In this study, several machine learning algorithms are applied for the prediction of the intracellular protein locations. These algorithms use the features extracted from protein sequences. In contrast, protein interactions have been less investigated.

Objectives

As protein interactions usually occur in the same or adjacent places, using this feature to find the location would be efficient and impressive. This study did not aim at increasing the total accuracy of the conducted research. The study has focused on the features of the proteins’ interaction and their employment which lead to a higher accuracy.

Materials and Methods

In this study, we have examined the protein interaction network as one of the features for prediction of the protein localization and its effects on the prediction results. In this regards, we have gathered some of the most common features including Amino Acid Composition, Dipeptide Compositions, Pseudo Amino Acid Compositions (PseAAC), Position Specific Scoring Matrix (PSSM), Functional Domain, Gene Ontology information, and the Pair-wise sequence alignment. The results of the classification are compared to the ones using protein interactions. For achieving this goal different machine learning algorithms were tested.

Results

The best-obtained results of using single feature set obtained using SVM classifier for PseAAC feature. The accuracy of combining all features with PPI data, using the Decision Tree and Random Forest classifiers, was 82.49% and 83.35%, respectively. In another experiment, using just protein interaction data with the different cutting points resulted in obtaining an accuracy of 93.035% for the protein location prediction.

Conclusion

In total, it was shown that protein(s) interaction has a significant impact on the prediction of the mitochondrial proteins’ location. This feature can separately distinguish the locations well. Using this feature the accuracy of the results is raised up to 5%.

Accumulation of potential driver genes with genomic alterations predicts survival of high-risk neuroblastoma patients

Background

Neuroblastoma is the most common pediatric malignancy with heterogeneous clinical behaviors, ranging from spontaneous regression to aggressive progression. Many studies have identified aberrations related to the pathogenesis and prognosis, broadly classifying neuroblastoma patients into high- and low-risk groups, but predicting tumor progression and clinical management of high-risk patients remains a big challenge.

Results

We integrate gene-level expression, array-based comparative genomic hybridization and functional gene-interaction network of 145 neuroblastoma patients to detect potential driver genes. The drivers are summarized into a driver-gene score (DGscore) for each patient, and we then validate its clinical relevance in terms of association with patient survival. Focusing on a subset of 48 clinically defined high-risk patients, we identify 193 recurrent regions of copy number alterations (CNAs), resulting in 274 altered genes whose copy-number gain or loss have parallel impact on the gene expression. Using a network enrichment analysis, we detect four common driver genes, ERCC6, HECTD2, KIAA1279, EMX2, and 66 patient-specific driver genes. Patients with high DGscore, thus carrying more copy-number-altered genes with correspondingly up- or down-regulated expression and functional implications, have worse survival than those with low DGscore (P = 0.006). Furthermore, Cox proportional-hazards regression analysis shows that, adjusted for age, tumor stage and MYCN amplification, DGscore is the only significant prognostic factor for high-risk neuroblastoma patients (P = 0.008).

Conclusions

Integration of genomic copy number alteration, expression and functional interaction-network data reveals clinically relevant and prognostic putative driver genes in high-risk neuroblastoma patients. The identified putative drivers are potential drug targets for individualized therapy.

Reviewers

This article was reviewed by Armand Valsesia, Susmita Datta and Aleksandra Gruca.

Electronic supplementary material

The online version of this article (10.1186/s13062-018-0218-5) contains supplementary material, which is available to authorized users.

Protein Sub-Nuclear Localization Based on Effective Fusion Representations and Dimension Reduction Algorithm LDA

An effective representation of a protein sequence plays a crucial role in protein sub-nuclear localization. The existing representations, such as dipeptide composition (DipC), pseudo-amino acid composition (PseAAC) and position specific scoring matrix (PSSM), are insufficient to represent protein sequence due to their single perspectives. Thus, this paper proposes two fusion feature representations of DipPSSM and PseAAPSSM to integrate PSSM with DipC and PseAAC, respectively. When constructing each fusion representation, we introduce the balance factors to value the importance of its components. The optimal values of the balance factors are sought by genetic algorithm. Due to the high dimensionality of the proposed representations, linear discriminant analysis (LDA) is used to find its important low dimensional structure, which is essential for classification and location prediction. The numerical experiments on two public datasets with KNN classifier and cross-validation tests showed that in terms of the common indexes of sensitivity, specificity, accuracy and MCC, the proposed fusing representations outperform the traditional representations in protein sub-nuclear localization, and the representation treated by LDA outperforms the untreated one.

Genome instability model of metastatic neuroblastoma tumorigenesis by a dictionary learning algorithm

Background

Metastatic neuroblastoma (NB) occurs in pediatric patients as stage 4S or stage 4 and it is characterized by heterogeneous clinical behavior associated with diverse genotypes. Tumors of stage 4 contain several structural copy number aberrations (CNAs) rarely found in stage 4S. To date, the NB tumorigenesis is not still elucidated, although it is evident that genomic instability plays a critical role in the genesis of the tumor. Here we propose a mathematical approach to decipher genomic data and we provide a new model of NB metastatic tumorigenesis.

Method

We elucidate NB tumorigenesis using Enhanced Fused Lasso Latent Feature Model (E-FLLat) modeling the array comparative chromosome hybridization (aCGH) data of 190 metastatic NBs (63 stage 4S and 127 stage 4). This model for aCGH segmentation, based on the minimization of functional dictionary learning (DL), combines several penalties tailored to the specificities of aCGH data. In DL, the original signal is approximated by a linear weighted combination of atoms: the elements of the learned dictionary.

Results

The hierarchical structures for stage 4S shows at the first level of the oncogenetic tree several whole chromosome gains except to the unbalanced gains of 17q, 2p and 2q. Conversely, the high CNA complexity found in stage 4 tumors, requires two different trees. Both stage 4 oncogenetic trees are marked diverged, up to five sublevels and the 17q gain is the most common event at the first level (2/3 nodes). Moreover the 11q deletion, one of the major unfavorable marker of disease progression, occurs before 3p loss indicating that critical chromosome aberrations appear at early stages of tumorigenesis. Finally, we also observed a significant (p = 0.025) association between patient age and chromosome loss in stage 4 cases.

Conclusion

These results led us to propose a genome instability progressive model in which NB cells initiate with a DNA synthesis uncoupled from cell division, that leads to stage 4S tumors, primarily characterized by numerical aberrations, or stage 4 tumors with high levels of genome instability resulting in complex chromosome rearrangements associated with high tumor aggressiveness and rapid disease progression.

Electronic supplementary material

The online version of this article (doi:10.1186/s12920-015-0132-y) contains supplementary material, which is available to authorized users.

Protein sub-nuclear localization prediction using SVM and Pfam domain information

The nucleus is the largest and the highly organized organelle of eukaryotic cells. Within nucleus exist a number of pseudo-compartments, which are not separated by any membrane, yet each of them contains only a specific set of proteins. Understanding protein sub-nuclear localization can hence be an important step towards understanding biological functions of the nucleus. Here we have described a method, SubNucPred developed by us for predicting the sub-nuclear localization of proteins. This method predicts protein localization for 10 different sub-nuclear locations sequentially by combining presence or absence of unique Pfam domain and amino acid composition based SVM model. The prediction accuracy during leave-one-out cross-validation for centromeric proteins was 85.05%, for chromosomal proteins 76.85%, for nuclear speckle proteins 81.27%, for nucleolar proteins 81.79%, for nuclear envelope proteins 79.37%, for nuclear matrix proteins 77.78%, for nucleoplasm proteins 76.98%, for nuclear pore complex proteins 88.89%, for PML body proteins 75.40% and for telomeric proteins it was 83.33%. Comparison with other reported methods showed that SubNucPred performs better than existing methods. A web-server for predicting protein sub-nuclear localization named SubNucPred has been established at http://14.139.227.92/mkumar/subnucpred/. Standalone version of SubNucPred can also be downloaded from the web-server.

Efficient and interpretable prediction of protein functional classes by correspondence analysis and compact set relations

Predicting protein functional classes such as localization sites and modifications plays a crucial role in function annotation. Given a tremendous amount of sequence data yielded from high-throughput sequencing experiments, the need of efficient and interpretable prediction strategies has been rapidly amplified. Our previous approach for subcellular localization prediction, PSLDoc, archives high overall accuracy for Gram-negative bacteria. However, PSLDoc is computational intensive due to incorporation of homology extension in feature extraction and probabilistic latent semantic analysis in feature reduction. Besides, prediction results generated by support vector machines are accurate but generally difficult to interpret.

In this work, we incorporate three new techniques to improve efficiency and interpretability. First, homology extension is performed against a compact non-redundant database using a fast search model to reduce running time. Second, correspondence analysis (CA) is incorporated as an efficient feature reduction to generate a clear visual separation of different protein classes. Finally, functional classes are predicted by a combination of accurate compact set (CS) relation and interpretable one-nearest neighbor (1-NN) algorithm. Besides localization data sets, we also apply a human protein kinase set to validate generality of our proposed method.

Experiment results demonstrate that our method make accurate prediction in a more efficient and interpretable manner. First, homology extension using a fast search on a compact database can greatly accelerate traditional running time up to twenty-five times faster without sacrificing prediction performance. This suggests that computational costs of many other predictors that also incorporate homology information can be largely reduced. In addition, CA can not only efficiently identify discriminative features but also provide a clear visualization of different functional classes. Moreover, predictions based on CS achieve 100% precision. When combined with 1-NN on unpredicted targets by CS, our method attains slightly better or comparable performance compared with the state-of-the-art systems.

SVM ensemble based transfer learning for large-scale membrane proteins discrimination

Membrane proteins play important roles in molecular trans-membrane transport, ligand-receptor recognition, cell-cell interaction, enzyme catalysis, host immune defense response and infectious disease pathways. Up to present, discriminating membrane proteins remains a challenging problem from the viewpoints of biological experimental determination and computational modeling. This work presents SVM ensemble based transfer learning model for membrane proteins discrimination (SVM-TLM). To reduce the data constraints on computational modeling, this method investigates the effectiveness of transferring the homolog knowledge to the target membrane proteins under the framework of probability weighted ensemble learning. As compared to multiple kernel learning based transfer learning model, the method takes the advantages of sparseness based SVM optimization on large data, thus more computationally efficient for large protein data analysis. The experiments on large membrane protein benchmark dataset show that SVM-TLM achieves significantly better cross validation performance than the baseline model.

Frequency of dipeptides and antidipeptides

Although it is reasonable to expect that the frequency of a generic dipeptide XY in proteins is the same of its counterpart YX, on the basis of an accurate statistical analysis of a large number of protein sequences, it appears that some dipeptides XY are considerably more frequent than their mirror images YX, referred to as antidipeptides. Given that it has been verified that this unexpected anisotropic frequency of occurrence is unbiased by the type of protein sequences that are analyzed, it is possible to conclude that this is a genuine phenomenon. Nevertheless, it was impossible to find the mechanism underlying this unexpected phenomenon, which does not seem to be related to diverse conformational propensities, to the different conformational flexibility of the peptide/antidipeptide pair, to dissimilar accessibility to the solvent or to gene random mutations.

Using over-represented tetrapeptides to predict protein submitochondria locations

The mitochondrion is a key organelle of eukaryotic cell that provides the energy for cellular activities. Correctly identifying submitochondria locations of proteins can provide plentiful information for understanding their functions. However, using web-experimental methods to recognize submitochondria locations of proteins are time-consuming and costly. Thus, it is highly desired to develop a bioinformatics method to predict the submitochondria locations of mitochondrion proteins. In this work, a novel method based on support vector machine was developed to predict the submitochondria locations of mitochondrion proteins by using over-represented tetrapeptides selected by using binomial distribution. A reliable and rigorous benchmark dataset including 495 mitochondrion proteins with sequence identity ≤25% was constructed for testing and evaluating the proposed model. Jackknife cross-validated results showed that the 91.1% of the 495 mitochondrion proteins can be correctly predicted. Subsequently, our model was estimated by three existing benchmark datasets. The overall accuracies are 94.0, 94.7 and 93.4%, respectively, suggesting that the proposed model is potentially useful in the realm of mitochondrion proteome research. Based on this model, we built a predictor called TetraMito which is freely available at http://lin.uestc.edu.cn/server/TetraMito.

An ensemble method for predicting subnuclear localizations from primary protein structures

Background

Predicting protein subnuclear localization is a challenging problem. Some previous works based on non-sequence information including Gene Ontology annotations and kernel fusion have respective limitations. The aim of this work is twofold: one is to propose a novel individual feature extraction method; another is to develop an ensemble method to improve prediction performance using comprehensive information represented in the form of high dimensional feature vector obtained by 11 feature extraction methods.

Methodology/Principal Findings

A novel two-stage multiclass support vector machine is proposed to predict protein subnuclear localizations. It only considers those feature extraction methods based on amino acid classifications and physicochemical properties. In order to speed up our system, an automatic search method for the kernel parameter is used. The prediction performance of our method is evaluated on four datasets: Lei dataset, multi-localization dataset, SNL9 dataset and a new independent dataset. The overall accuracy of prediction for 6 localizations on Lei dataset is 75.2% and that for 9 localizations on SNL9 dataset is 72.1% in the leave-one-out cross validation, 71.7% for the multi-localization dataset and 69.8% for the new independent dataset, respectively. Comparisons with those existing methods show that our method performs better for both single-localization and multi-localization proteins and achieves more balanced sensitivities and specificities on large-size and small-size subcellular localizations. The overall accuracy improvements are 4.0% and 4.7% for single-localization proteins and 6.5% for multi-localization proteins. The reliability and stability of our classification model are further confirmed by permutation analysis.

Conclusions

It can be concluded that our method is effective and valuable for predicting protein subnuclear localizations. A web server has been designed to implement the proposed method. It is freely available at http://bioinformatics.awowshop.com/snlpred_page.php.

Copy number variation signature to predict human ancestry

BACKGROUND

Copy number variations (CNVs) are genomic structural variants that are found in healthy populations and have been observed to be associated with disease susceptibility. Existing methods for CNV detection are often performed on a sample-by-sample basis, which is not ideal for large datasets where common CNVs must be estimated by comparing the frequency of CNVs in the individual samples. Here we describe a simple and novel approach to locate genome-wide CNVs common to a specific population, using human ancestry as the phenotype.

RESULTS

We utilized our previously published Genome Alteration Detection Analysis (GADA) algorithm to identify common ancestry CNVs (caCNVs) and built a caCNV model to predict population structure. We identified a 73 caCNV signature using a training set of 225 healthy individuals from European, Asian, and African ancestry. The signature was validated on an independent test set of 300 individuals with similar ancestral background. The error rate in predicting ancestry in this test set was 2% using the 73 caCNV signature. Among the caCNVs identified, several were previously confirmed experimentally to vary by ancestry. Our signature also contains a caCNV region with a single microRNA (MIR270), which represents the first reported variation of microRNA by ancestry.

CONCLUSIONS

We developed a new methodology to identify common CNVs and demonstrated its performance by building a caCNV signature to predict human ancestry with high accuracy. The utility of our approach could be extended to large case-control studies to identify CNV signatures for other phenotypes such as disease susceptibility and drug response.

Interrogating noise in protein sequences from the perspective of protein-protein interactions prediction

The past decades witnessed extensive efforts to study the relationship among proteins. Particularly, sequence-based protein-protein interactions (PPIs) prediction is fundamentally important in speeding up the process of mapping interactomes of organisms. High-throughput experimental methodologies make many model organism's PPIs known, which allows us to apply machine learning methods to learn understandable rules from the available PPIs. Under the machine learning framework, the composition vectors are usually applied to encode proteins as real-value vectors. However, the composition vector value might be highly correlated to the distribution of amino acids, i.e., amino acids which are frequently observed in nature tend to have a large value of composition vectors. Thus formulation to estimate the noise induced by the background distribution of amino acids may be needed during representations. Here, we introduce two kinds of denoising composition vectors, which were successfully used in construction of phylogenetic trees, to eliminate the noise. When validating these two denoising composition vectors on Escherichia coli (E. coli), Saccharomyces cerevisiae (S. cerevisiae) and human PPIs datasets, surprisingly, the predictive performance is not improved, and even worse than non-denoised prediction. These results suggest that the noise in phylogenetic tree construction may be valuable information in PPIs prediction.

Predicting secretory proteins of malaria parasite by incorporating sequence evolution information into pseudo amino acid composition via grey system model

The malaria disease has become a cause of poverty and a major hindrance to economic development. The culprit of the disease is the parasite, which secretes an array of proteins within the host erythrocyte to facilitate its own survival. Accordingly, the secretory proteins of malaria parasite have become a logical target for drug design against malaria. Unfortunately, with the increasing resistance to the drugs thus developed, the situation has become more complicated. To cope with the drug resistance problem, one strategy is to timely identify the secreted proteins by malaria parasite, which can serve as potential drug targets. However, it is both expensive and time-consuming to identify the secretory proteins of malaria parasite by experiments alone. To expedite the process for developing effective drugs against malaria, a computational predictor called "iSMP-Grey" was developed that can be used to identify the secretory proteins of malaria parasite based on the protein sequence information alone. During the prediction process a protein sample was formulated with a 60D (dimensional) feature vector formed by incorporating the sequence evolution information into the general form of PseAAC (pseudo amino acid composition) via a grey system model, which is particularly useful for solving complicated problems that are lack of sufficient information or need to process uncertain information. It was observed by the jackknife test that iSMP-Grey achieved an overall success rate of 94.8%, remarkably higher than those by the existing predictors in this area. As a user-friendly web-server, iSMP-Grey is freely accessible to the public at http://www.jci-bioinfo.cn/iSMP-Grey. Moreover, for the convenience of most experimental scientists, a step-by-step guide is provided on how to use the web-server to get the desired results without the need to follow the complicated mathematical equations involved in this paper.

Multi-label multi-kernel transfer learning for human protein subcellular localization

Recent years have witnessed much progress in computational modelling for protein subcellular localization. However, the existing sequence-based predictive models demonstrate moderate or unsatisfactory performance, and the gene ontology (GO) based models may take the risk of performance overestimation for novel proteins. Furthermore, many human proteins have multiple subcellular locations, which renders the computational modelling more complicated. Up to the present, there are far few researches specialized for predicting the subcellular localization of human proteins that may reside in multiple cellular compartments. In this paper, we propose a multi-label multi-kernel transfer learning model for human protein subcellular localization (MLMK-TLM). MLMK-TLM proposes a multi-label confusion matrix, formally formulates three multi-labelling performance measures and adapts one-against-all multi-class probabilistic outputs to multi-label learning scenario, based on which to further extends our published work GO-TLM (gene ontology based transfer learning model for protein subcellular localization) and MK-TLM (multi-kernel transfer learning based on Chou's PseAAC formulation for protein submitochondria localization) for multiplex human protein subcellular localization. With the advantages of proper homolog knowledge transfer, comprehensive survey of model performance for novel protein and multi-labelling capability, MLMK-TLM will gain more practical applicability. The experiments on human protein benchmark dataset show that MLMK-TLM significantly outperforms the baseline model and demonstrates good multi-labelling ability for novel human proteins. Some findings (predictions) are validated by the latest Swiss-Prot database. The software can be freely downloaded at http://soft.synu.edu.cn/upload/msy.rar.

Regions of homozygosity in three Southeast Asian populations

The genomes of outbred populations were first shown in 2006 to contain regions of homozygosity (ROHs) of several megabases. Further studies have also investigated the characteristics of ROHs in healthy individuals in various populations but there are no studies on Singapore populations to date. This study aims to identify and investigate the characteristics of ROHs in three Singapore populations. A total of 268 samples (96 Chinese, 89 Malays and 83 Indians) are genotyped on Illumina Human 1 M Beadchip and Affymetrix Genome-Wide Human SNP Array 6.0. We use the PennCNV algorithm to detect ROHs. We report an abundance of ROHs (≥500 kb), with an average of more than one hundred regions per individual. On average, the Indian population has the lowest number of ROHs and smallest total length of ROHs per individual compared with the Chinese and Malay populations. We further investigate the relationship between the occurrence of ROHs and haplotype frequency, regional linkage disequilibrium (LD) and positive selection. Based on the results of this data set, we find that the frequency of occurrence of ROHs is positively associated with haplotype frequency and regional LD. The majority of regions detected for recent positive selection and regions with differential LD between populations overlap with the ROH loci. When we consider both the location of the ROHs and the allelic form of the ROHs, we are able to separate the populations by principal component analysis, demonstrating that ROHs contain information on population structure and the demographic history of a population.

Recent progress in predicting protein sub-subcellular locations

In the last two decades, the number of the known protein sequences increased very rapidly. However, a knowledge of protein function only exists for a small portion of these sequences. Since the experimental approaches for determining protein functions are costly and time consuming, in silico methods have been introduced to bridge the gap between knowledge of protein sequences and their functions. Knowing the subcellular location of a protein is considered to be a critical step in understanding its biological functions. Many efforts have been undertaken to predict the protein subcellular locations in silico. With the accumulation of available data, the substructures of some subcellular organelles, such as the cell nucleus, mitochondria and chloroplasts, have been taken into consideration by several studies in recent years. These studies create a new research topic, namely 'protein sub-subcellular location prediction', which goes one level deeper than classic protein subcellular location prediction.

Copy number polymorphisms in new HapMap III and Singapore populations

Copy number variations can be identified using newer genotyping arrays with higher single nucleotide polymorphisms (SNPs) density and copy number probes accompanied by newer algorithms. McCarroll et al. (2008) applied these to the HapMap II samples and identified 1316 copy number polymorphisms (CNPs). In our study, we applied the same approach to 859 samples from three Singapore populations and seven HapMap III populations. Approximately 50% of the 1291 autosomal CNPs were found to be polymorphic only in populations of non-African ancestry. Pairwise comparisons among the 10 populations showed substantial differences in the CNPs frequencies. Additionally, 698 CNPs showed significant differences with false discovery rate (FDR)<0.01 among the 10 populations and these loci overlap with known disease-associated or pharmacogenetic-related genes such as CFHR3 and CFHR1 (age related macular degeneration), GSTTI (metabolism of various carcinogenic compounds and cancers) and UGT2B17 (prostate cancer and graft-versus-host disease). The correlations between CNPs and genome-wide association studies-SNPs were investigated and several loci, which were previously unreported, that may potentially be implicated in complex diseases and traits were found; for example, childhood acute lymphoblastic leukaemia, age-related macular degeneration, breast cancer, response to antipsychotic treatment, rheumatoid arthritis and type-1 diabetes. Additionally, we also found 5014 novel copy number loci that have not been reported previously by McCarroll et al. (2008) in the 10 populations.

A population-based study of copy number variants and regions of homozygosity in healthy Swedish individuals

The abundance of copy number variants (CNVs) and regions of homozygosity (ROHs) have been well documented in previous studies. In addition, their roles in complex diseases and traits have since been increasingly appreciated. However, only a limited amount of CNV and ROH data is currently available for the Swedish population. We conducted a population-based study to detect and characterize CNVs and ROHs in 87 randomly selected healthy Swedish individuals using the Affymetrix SNP Array 6.0. More than 600 CNV loci were detected in the population using two different CNV-detection algorithms (PennCNV and Birdsuite). A total of 196 loci were consistently identified by both algorithms, suggesting their reliability. Numerous disease-associated and pharmacogenetics-related genes were found to be overlapping with common CNV loci such as CFHR1/R3, LCE3B/3C, UGT2B17 and GSTT1. Correlation analysis between copy number polymorphisms (CNPs) and genome-wide association studies-identified single-nucleotide polymorphisms also indicates the potential roles of several CNPs as causal variants for diseases and traits such as body mass index, Crohn's disease and multiple sclerosis. In addition, we also identified a total of 14 815 ROHs 500 kb or 2814 ROHs 1M in the Swedish individuals with an average of 170 and 32 regions detected per individual respectively. Approximately 141 Mb or 4.92% of the genome is homozygous in each individual of the Swedish population. This is the first population-based study to investigate the population characteristics of CNVs and ROHs in the Swedish population. This study found many CNV loci that warrant further investigation, and also highlighted the abundance and importance of investigating ROHs for their associations with complex diseases and traits.

Gene ontology based transfer learning for protein subcellular localization

BACKGROUND

Prediction of protein subcellular localization generally involves many complex factors, and using only one or two aspects of data information may not tell the true story. For this reason, some recent predictive models are deliberately designed to integrate multiple heterogeneous data sources for exploiting multi-aspect protein feature information. Gene ontology, hereinafter referred to as GO, uses a controlled vocabulary to depict biological molecules or gene products in terms of biological process, molecular function and cellular component. With the rapid expansion of annotated protein sequences, gene ontology has become a general protein feature that can be used to construct predictive models in computational biology. Existing models generally either concatenated the GO terms into a flat binary vector or applied majority-vote based ensemble learning for protein subcellular localization, both of which can not estimate the individual discriminative abilities of the three aspects of gene ontology.

RESULTS

In this paper, we propose a Gene Ontology Based Transfer Learning Model (GO-TLM) for large-scale protein subcellular localization. The model transfers the signature-based homologous GO terms to the target proteins, and further constructs a reliable learning system to reduce the adverse affect of the potential false GO terms that are resulted from evolutionary divergence. We derive three GO kernels from the three aspects of gene ontology to measure the GO similarity of two proteins, and derive two other spectrum kernels to measure the similarity of two protein sequences. We use simple non-parametric cross validation to explicitly weigh the discriminative abilities of the five kernels, such that the time & space computational complexities are greatly reduced when compared to the complicated semi-definite programming and semi-indefinite linear programming. The five kernels are then linearly merged into one single kernel for protein subcellular localization. We evaluate GO-TLM performance against three baseline models: MultiLoc, MultiLoc-GO and Euk-mPLoc on the benchmark datasets the baseline models adopted. 5-fold cross validation experiments show that GO-TLM achieves substantial accuracy improvement against the baseline models: 80.38% against model Euk-mPLoc 67.40% with 12.98% substantial increase; 96.65% and 96.27% against model MultiLoc-GO 89.60% and 89.60%, with 7.05% and 6.67% accuracy increase on dataset MultiLoc plant and dataset MultiLoc animal, respectively; 97.14%, 95.90% and 96.85% against model MultiLoc-GO 83.70%, 90.10% and 85.70%, with accuracy increase 13.44%, 5.8% and 11.15% on dataset BaCelLoc plant, dataset BaCelLoc fungi and dataset BaCelLoc animal respectively. For BaCelLoc independent sets, GO-TLM achieves 81.25%, 80.45% and 79.46% on dataset BaCelLoc plant holdout, dataset BaCelLoc plant holdout and dataset BaCelLoc animal holdout, respectively, as compared against baseline model MultiLoc-GO 76%, 60.00% and 73.00%, with accuracy increase 5.25%, 20.45% and 6.46%, respectively.

CONCLUSIONS

Since direct homology-based GO term transfer may be prone to introducing noise and outliers to the target protein, we design an explicitly weighted kernel learning system (called Gene Ontology Based Transfer Learning Model, GO-TLM) to transfer to the target protein the known knowledge about related homologous proteins, which can reduce the risk of outliers and share knowledge between homologous proteins, and thus achieve better predictive performance for protein subcellular localization. Cross validation and independent test experimental results show that the homology-based GO term transfer and explicitly weighing the GO kernels substantially improve the prediction performance.