Isis Bonet

Predicting Human Immunodeficiency Virus (HIV) Drug Resistance Using Recurrent Neural Networks

Predicting HIV resistance to drugs is one of many problems for which bioinformaticians have implemented and trained machine learning methods, such as neural networks. Predicting HIV resistance would be much easier if we could directly use the three-dimensional (3D) structure of the targeted protein sequences, but unfortunately we rarely have enough structural information available to train a neural network. Fur-thermore, prediction of the 3D structure of a protein is not straightforward. However, characteristics related to the 3D structure can be used to train a machine learning algorithm as an alternative to take into account the information of the protein folding in the 3D space. Here, starting from this philosophy, we select the amino acid energies as features to predict HIV drug resistance, using a specific topology of a neural network. In this paper, we demonstrate that the amino acid ener-gies are good features to represent the HIV genotype. In addi-tion, it was shown that Bidirectional Recurrent Neural Networks can be used as an efficient classification method for this prob-lem. The prediction performance that was obtained was greater than or at least comparable to results obtained previously. The accuracies vary between 81.3% and 94.7%.

Comparing distance measures with visual methods

The selection of the distance measure to separate the objects of the knowledge space is critical in many classification algorithms. In this paper, we analyze the distance measures reported in the literature for the problem of HIV prediction. We propose a new distance for HIV viral sequences, based on the mutations with regard to the HXB2 reference sequence. In a first step, we reduce data dimensionality in order to subsequently analyze the distance measures performance in terms of its ability to separate classes.

Feature extraction using clustering of protein

In this paper we investigate the usage of a clustering algorithm as a feature extraction technique to find new features to represent the protein sequence. In particular, our work focuses on the prediction of HIV protease resistance to drugs. We use a biologically motivated similarity function based on the contact energy of the amino acid and the position in the sequence. The performance measure was computed taking into account the clustering reliability and the classification validity. An SVM using 10-fold crossvalidation and the k-means algorithm were used for classification and clustering respectively. The best results were obtained by reducing an initial set of 99 features to a lower dimensional feature set of 36-66 features.

Iterative Clustering Method for Metagenomic Sequences

Metagenomics studies microbial DNA of environmental samples. The sequencing tools produce a set of genome fragments providing a challenge for metagenomics to associate them with the corresponding phylogenetic group. To solve this problem there are binning methods, which are classified into two sequencing categories: similarity and composition. This paper proposes an iterative clustering method, which aim at achieving a low sensitivity of clusters. The approach consists of iteratively run k-means reducing the training data in each step. Selection of data for next iteration depends on the result obtained in the previous, which is based on the compactness measure. The final performance clustering is evaluated according with the sensitivity of clusters. The results demonstrate that proposed model is better than the simple k-means for metagenome databases.

Multi-Classifier Based on Hard Instances- New Method for Prediction of Human Immunodeficiency Virus Drug Resistance

There are several classification problems, which are difficult to solve using a single classifier because of the complexity of the decision boundary. Whereas a wide variety of multiple classifier systems have been built with the purpose of improving the recognition process, there is no universal method performing the best. This paper provides a review of different multi-classifiers and some application of them. Also it is shown a novel model of combining classifiers and its application to predicting human immunodeficiency virus drug resistance from genotype. The proposal is based on the use of different classifier models. It clusters the dataset considering the performance of the base classifiers. The system learns how to decide from the groups, by using a meta-classifier, which are the best classifiers for a given pattern. The proposed model is compared with well-known classifier ensembles and individual classifiers as well resulting the novel model in similar or even better performance.

Comparing Distance Measures with Visual Methods

The selection of the distance measure to separate the objects of the knowledge space is critical in many classification algorithms. In this paper, we analyze the distance measures reported in the literature for the problem of HIV prediction. We propose a new distance for HIV viral sequences, based on the mutations with regard to the HXB2 reference sequence. In a first step, we reduce data dimensionality in order to subsequently analyze the distance measures performance in terms of its ability to separate classes.

Boolean algebraic structures of the genetic code: possibilities of applications

Authors had reported before two dual Boolean algebras to understand the underlying logic of the genetic code structure. In such Boolean structures, deductions have physico-chemical meaning. We summarize here that these algebraic structures can help us to describe the gene evolution process. Particularly in the experimental confrontation, it was found that most of the mutations of several proteins correspond to deductions in these algebras and they have a small Hamming distance related to their respective wild type. Two applications of the corresponding codification system in bioinformatics problems are also shown. The first one is the classification of mutations in a protein. The other one is related with the problem of detecting donors and acceptors in DNA sequences. Besides, pure mathematical models, Statistical techniques (Decision Trees) and Artificial Intelligence techniques (Bayesian Networks) were used in order to show how to accomplish them to solve these knowledge-discovery practical problems.

Consensus Clustering for Binning Metagenome Sequences

The advances in next-generation sequencing technologies allow researchers to sequence in parallel millions of microbial organisms directly from environmental samples. The result of this “shotgun” sequencing are many short DNA fragments of different organisms, which constitute the basis for the field of metagenomics. Although there are big databases with known microbial DNA that allow us classify some fragments, these databases only represent around 1% of all the species existing in the entire world. For this reason, it is important to use unsupervised methods to group the fragments with the same taxonomic levels. In this paper we focus on the binning step in metagenomics in an unsupervised way. We propose a consensus clustering method based on an iterative clustering process using different lengths of sequences in the databases and a mixture of distance as approach to finding the consensus clustering. The final performance clustering is evaluated according with the purity of clusters. The results achieved by the proposed method outperforms results obtained by simple methods and iterative methods.

Bidirectional Recurrent Neural Networks for Biological Sequences Prediction

The aim of this paper is to analyze the potentialities of Bidirectional Recurrent Neural Networks in classification problems. Different functions are proposed to merge the network outputs into one single classification decision. In order to analyze when these networks could be useful; artificial datasets were constructed to compare their performance against well-known classification methods in different situations, such as complex and simple decision boundaries, and related and independent features. The advantage of this neural network in classification problems with complicated decision boundaries and feature relations was proved statistically. Finally, better results using this network topology in the prediction of HIV drug resistance were also obtained.

Ensemble of classifiers based on hard instances

There are several classification problems, which are difficult to solve using a single classifier because of the complexity of the decision boundary. Whereas, a wide variety of multiple classifier systems have been built with the purpose of improving the recognition process. There is no universal method performing the best. The aim of this paper is to show another model of combining classifiers. This model is based on the use of different classifier models. It makes clusters to divide the dataset, taking into account the performance of the base classifiers. The system learns how to decide from the groups, by a meta-classifier, who are the best classifiers for a given pattern. In order to compare the new model with well-known classifier ensembles, we carried out experiments with some international databases. The results demonstrate that this new model can achieve similar or better performance than the classic ensembles.