Benchmarking protein classification algorithms via supervised cross-validation

LZW-Kernel: fast kernel utilizing variable length code blocks from LZW compressors for protein sequence classification

Motivation

Bioinformatics studies often rely on similarity measures between sequence pairs, which often pose a bottleneck in large-scale sequence analysis.

Results

Here, we present a new convolutional kernel function for protein sequences called the Lempel-Ziv-Welch (LZW)-Kernel. It is based on code words identified with the LZW universal text compressor. The LZW-Kernel is an alignment-free method, it is always symmetric, is positive, always provides 1.0 for self-similarity and it can directly be used with Support Vector Machines (SVMs) in classification problems, contrary to normalized compression distance, which often violates the distance metric properties in practice and requires further techniques to be used with SVMs. The LZW-Kernel is a one-pass algorithm, which makes it particularly plausible for big data applications. Our experimental studies on remote protein homology detection and protein classification tasks reveal that the LZW-Kernel closely approaches the performance of the Local Alignment Kernel (LAK) and the SVM-pairwise method combined with Smith-Waterman (SW) scoring at a fraction of the time. Moreover, the LZW-Kernel outperforms the SVM-pairwise method when combined with Basic Local Alignment Search Tool (BLAST) scores, which indicates that the LZW code words might be a better basis for similarity measures than local alignment approximations found with BLAST. In addition, the LZW-Kernel outperforms n-gram based mismatch kernels, hidden Markov model based SAM and Fisher kernel and protein family based PSI-BLAST, among others. Further advantages include the LZW-Kernel's reliance on a simple idea, its ease of implementation, and its high speed, three times faster than BLAST and several magnitudes faster than SW or LAK in our tests.

Availability and implementation

LZW-Kernel is implemented as a standalone C code and is a free open-source program distributed under GPLv3 license and can be downloaded from https://github.com/kfattila/LZW-Kernel.

Supplementary information

Supplementary data are available at Bioinformatics Online.

GaIn: Human Gait Inference for Lower Limbic Prostheses for Patients Suffering from Double Trans-Femoral Amputation

Several studies have analyzed human gait data obtained from inertial gyroscope and accelerometer sensors mounted on different parts of the body. In this article, we take a step further in gait analysis and provide a methodology for predicting the movements of the legs, which can be applied in prosthesis to imitate the missing part of the leg in walking. In particular, we propose a method, called GaIn, to control non-invasive, robotic, prosthetic legs. GaIn can infer the movements of both missing shanks and feet for humans suffering from double trans-femoral amputation using biologically inspired recurrent neural networks. Predictions are performed for casual walking related activities such as walking, taking stairs, and running based on thigh movement. In our experimental tests, GaIn achieved a 4.55 prediction error for shank movements on average. However, a patient's intention to stand up and sit down cannot be inferred from thigh movements. In fact, intention causes thigh movements while the shanks and feet remain roughly still. The GaIn system can be triggered by thigh muscle activities measured with electromyography (EMG) sensors to make robotic prosthetic legs perform standing up and sitting down actions. The GaIn system has low prediction latency and is fast and computationally inexpensive to be deployed on mobile platforms and portable devices.

Automatic recognition of gait patterns in human motor disorders using machine learning: A review

BACKGROUND

automatic recognition of human movement is an effective strategy to assess abnormal gait patterns. Machine learning approaches are mainly applied due to their ability to work with multidimensional nonlinear features.

PURPOSE

to compare several machine learning algorithms employed for gait pattern recognition in motor disorders using discriminant features extracted from gait dynamics. Additionally, this work highlights procedures that improve gait recognition performance.

METHODS

we conducted an electronic literature search on Web of Science, IEEE, and Scopus, using "human recognition", "gait patterns'', and "feature selection methods" as relevant keywords.

RESULTS

analysis of the literature showed that kernel principal component analysis and genetic algorithms are efficient at reducing dimensional features due to their ability to process nonlinear data and converge to global optimum. Comparative analysis of machine learning performance showed that support vector machines (SVMs) exhibited higher accuracy and proper generalization for new instances.

CONCLUSIONS

automatic recognition by combining dimensional data reduction, cross-validation and normalization techniques with SVMs may offer an objective and rapid tool for investigating the subject's clinical status. Future directions comprise the real-time application of these tools to drive powered assistive devices in free-living conditions.

CX, DPX, and PCW: Web Servers for the Visualization of Interior and Protruding Regions of Protein Structures in 3D and 1D

The CX and DPX web-based servers at http://pongor.itk.ppke.hu/bioinfoservices are dedicated to the analysis of protein 3D structures submitted by the users as Protein Data Bank (PDB) files. CX computes an atomic protrusion index, cx that makes it possible to highlight the protruding atoms within a protein 3D structure. DPX calculates a depth index, dpx for buried atoms, and allows one to visualize the distribution of buried residues. CX and DPX visualize 3D structures colored according to the calculated indices and return PDB files that can be visualized using standard programs. A combined server site, the Protein Core Workbench allows visualization of dpx, cx, solvent-accessible area as well as the number of atomic contacts as 3D plots and 1D sequence plots. Online visualization of the 3D structures and 1D sequence plots are available in all three servers. Mirror sites are available at http://hydra.icgeb.trieste.it/protein/ .

Crop classification by forward neural network with adaptive chaotic particle swarm optimization

This paper proposes a hybrid crop classifier for polarimetric synthetic aperture radar (SAR) images. The feature sets consisted of span image, the H/A/α decomposition, and the gray-level co-occurrence matrix (GLCM) based texture features. Then, the features were reduced by principle component analysis (PCA). Finally, a two-hidden-layer forward neural network (NN) was constructed and trained by adaptive chaotic particle swarm optimization (ACPSO). K-fold cross validation was employed to enhance generation. The experimental results on Flevoland sites demonstrate the superiority of ACPSO to back-propagation (BP), adaptive BP (ABP), momentum BP (MBP), Particle Swarm Optimization (PSO), and Resilient back-propagation (RPROP) methods. Moreover, the computation time for each pixel is only 1.08 × 10⁻⁷ s.

Computational approaches for protein function prediction: a combined strategy from multiple sequence alignment to molecular docking-based virtual screening

The functional characterization of proteins represents a daily challenge for biochemical, medical and computational sciences. Although finally proved on the bench, the function of a protein can be successfully predicted by computational approaches that drive the further experimental assays. Current methods for comparative modeling allow the construction of accurate 3D models for proteins of unknown structure, provided that a crystal structure of a homologous protein is available. Binding regions can be proposed by using binding site predictors, data inferred from homologous crystal structures, and data provided from a careful interpretation of the multiple sequence alignment of the investigated protein and its homologs. Once the location of a binding site has been proposed, chemical ligands that have a high likelihood of binding can be identified by using ligand docking and structure-based virtual screening of chemical libraries. Most docking algorithms allow building a list sorted by energy of the lowest energy docking configuration for each ligand of the library. In this review the state-of-the-art of computational approaches in 3D protein comparative modeling and in the study of protein-ligand interactions is provided. Furthermore a possible combined/concerted multistep strategy for protein function prediction, based on multiple sequence alignment, comparative modeling, binding region prediction, and structure-based virtual screening of chemical libraries, is described by using suitable examples. As practical examples, Abl-kinase molecular modeling studies, HPV-E6 protein multiple sequence alignment analysis, and some other model docking-based characterization reports are briefly described to highlight the importance of computational approaches in protein function prediction.

Balanced ROC analysis (BAROC) protocol for the evaluation of protein similarities

Identification of problematic protein classes (domain types, protein families) that are difficult to predict from sequence is a key issue in genome annotation. ROC (Receiver Operating Characteristic) analysis is routinely used for the evaluation of protein similarities, however its results - the area under curve (AUC) values - are differentially biased for the various protein classes that are highly different in size. We show the bias can be compensated for by adjusting the length of the top list in a class-dependent fashion, so that the number of negatives within the top list will be equal to (or proportional with) the size of the positive class. Using this balanced protocol the problematic classes can be identified by their AUC values, or by a scatter diagram in which the AUC values are plotted against positive/negative ratio of the top list. The use of likelihood-ratio scoring (Kaján et al, Bioinformatics,22, 2865-2869, 2007) the bias caused by class imbalance can be further decreased.