Predicting human protein subcellular localization by heterogeneous and comprehensive approaches

TripletProt: Deep Representation Learning of Proteins Based On Siamese Networks

Pretrained representations have recently gained attention in various machine learning applications. Nonetheless, the high computational costs associated with training these models have motivated alternative approaches for representation learning. Herein we introduce TripletProt, a new approach for protein representation learning based on the Siamese neural networks. Representation learning of biological entities which capture essential features can alleviate many of the challenges associated with supervised learning in bioinformatics. The most important distinction of our proposed method is relying on the protein-protein interaction (PPI) network. The computational cost of the generated representations for any potential application is significantly lower than comparable methods since the length of the representations is significantly smaller than that in other approaches. TripletProt offers great potentials for the protein informatics tasks and can be widely applied to similar tasks. We evaluate TripletProt comprehensively in protein functional annotation tasks including sub-cellular localization (14 categories) and gene ontology prediction (more than 2000 classes), which are both challenging multi-class, multi-label classification machine learning problems. We compare the performance of TripletProt with the state-of-the-art approaches including a recurrent language model-based approach (i.e., UniRep), as well as a protein-protein interaction (PPI) network and sequence-based method (i.e., DeepGO). Our TripletProt showed an overall improvement of F1 score in the above mentioned comprehensive functional annotation tasks, solely relying on the PPI network. Availability: The source code and datasets are available at https://github.com/EsmaeilNourani/TripletProt.

Neural Network and Random Forest Models in Protein Function Prediction

Over the past decade, the demand for automated protein function prediction has increased due to the volume of newly sequenced proteins. In this paper, we address the function prediction task by developing an ensemble system automatically assigning Gene Ontology (GO) terms to the given input protein sequence. We develop an ensemble system which combines the GO predictions made by random forest (RF) and neural network (NN) classifiers. Both RF and NN models rely on features derived from BLAST sequence alignments, taxonomy and protein signature analysis tools. In addition, we report on experiments with a NN model that directly analyzes the amino acid sequence as its sole input, using a convolutional layer. The Swiss-Prot database is used as the training and evaluation data. In the CAFA3 evaluation, which relies on experimental verification of the functional predictions, our submitted ensemble model demonstrates competitive performance ranking among top-10 best-performing systems out of over 100 submitted systems. In this paper, we evaluate and further improve the CAFA3-submitted system. Our machine learning models together with the data pre-processing and feature generation tools are publicly available as an open source software at https://github.com/TurkuNLP/CAFA3.

Ensemble of Multiple Classifiers for Multilabel Classification of Plant Protein Subcellular Localization

The accurate prediction of protein localization is a critical step in any functional genome annotation process. This paper proposes an improved strategy for protein subcellular localization prediction in plants based on multiple classifiers, to improve prediction results in terms of both accuracy and reliability. The prediction of plant protein subcellular localization is challenging because the underlying problem is not only a multiclass, but also a multilabel problem. Generally, plant proteins can be found in 10–14 locations/compartments. The number of proteins in some compartments (nucleus, cytoplasm, and mitochondria) is generally much greater than that in other compartments (vacuole, peroxisome, Golgi, and cell wall). Therefore, the problem of imbalanced data usually arises. Therefore, we propose an ensemble machine learning method based on average voting among heterogeneous classifiers. We first extracted various types of features suitable for each type of protein localization to form a total of 479 feature spaces. Then, feature selection methods were used to reduce the dimensions of the features into smaller informative feature subsets. This reduced feature subset was then used to train/build three different individual models. In the process of combining the three distinct classifier models, we used an average voting approach to combine the results of these three different classifiers that we constructed to return the final probability prediction. The method could predict subcellular localizations in both single- and multilabel locations, based on the voting probability. Experimental results indicated that the proposed ensemble method could achieve correct classification with an overall accuracy of 84.58% for 11 compartments, on the basis of the testing dataset.

Plant-mSubP: a computational framework for the prediction of single- and multi-target protein subcellular localization using integrated machine-learning approaches

The subcellular localization of proteins is very important for characterizing its function in a cell. Accurate prediction of the subcellular locations in computational paradigm has been an active area of interest. Most of the work has been focused on single localization prediction. Only few studies have discussed the multi-target localization, but have not achieved good accuracy so far; in plant sciences, very limited work has been done. Here we report the development of a novel tool Plant-mSubP, which is based on integrated machine learning approaches to efficiently predict the subcellular localizations in plant proteomes. The proposed approach predicts with high accuracy 11 single localizations and three dual locations of plant cell. Several hybrid features based on composition and physicochemical properties of a protein such as amino acid composition, pseudo amino acid composition, auto-correlation descriptors, quasi-sequence-order descriptors and hybrid features are used to represent the protein. The performance of the proposed method has been assessed through a training set as well as an independent test set. Using the hybrid feature of the pseudo amino acid composition, N-Center-C terminal amino acid composition and the dipeptide composition (PseAAC-NCC-DIPEP), an overall accuracy of 81.97 %, 84.75 % and 87.88 % is achieved on the training data set of proteins containing the single-label, single- and dual-label combined, and dual-label proteins, respectively. When tested on the independent data, an accuracy of 64.36 %, 64.84 % and 81.08 % is achieved on the single-label, single- and dual-label, and dual-label proteins, respectively. The prediction models have been implemented on a web server available at http://bioinfo.usu.edu/Plant-mSubP/. The results indicate that the proposed approach is comparable to the existing methods in single localization prediction and outperforms all other existing tools when compared for dual-label proteins. The prediction tool will be a useful resource for better annotation of various plant proteomes.

Machine learning techniques for protein function prediction

Proteins play important roles in living organisms, and their function is directly linked with their structure. Due to the growing gap between the number of proteins being discovered and their functional characterization (in particular as a result of experimental limitations), reliable prediction of protein function through computational means has become crucial. This paper reviews the machine learning techniques used in the literature, following their evolution from simple algorithms such as logistic regression to more advanced methods like support vector machines and modern deep neural networks. Hyperparameter optimization methods adopted to boost prediction performance are presented. In parallel, the metamorphosis in the features used by these algorithms from classical physicochemical properties and amino acid composition, up to text-derived features from biomedical literature and learned feature representations using autoencoders, together with feature selection and dimensionality reduction techniques, are also reviewed. The success stories in the application of these techniques to both general and specific protein function prediction are discussed.

Prediction of bacterial E3 ubiquitin ligase effectors using reduced amino acid peptide fingerprinting

Background

Although pathogenic Gram-negative bacteria lack their own ubiquitination machinery, they have evolved or acquired virulence effectors that can manipulate the host ubiquitination process through structural and/or functional mimicry of host machinery. Many such effectors have been identified in a wide variety of bacterial pathogens that share little sequence similarity amongst themselves or with eukaryotic ubiquitin E3 ligases.

Methods

To allow identification of novel bacterial E3 ubiquitin ligase effectors from protein sequences we have developed a machine learning approach, the SVM-based Identification and Evaluation of Virulence Effector Ubiquitin ligases (SIEVE-Ub). We extend the string kernel approach used previously to sequence classification by introducing reduced amino acid (RED) alphabet encoding for protein sequences.

Results

We found that 14mer peptides with amino acids represented as simply either hydrophobic or hydrophilic provided the best models for discrimination of E3 ligases from other effector proteins with a receiver-operator characteristic area under the curve (AUC) of 0.90. When considering a subset of E3 ubiquitin ligase effectors that do not fall into known sequence based families we found that the AUC was 0.82, demonstrating the effectiveness of our method at identifying novel functional family members. Feature selection was used to identify a parsimonious set of 10 RED peptides that provided good discrimination, and these peptides were found to be located in functionally important regions of the proteins involved in E2 and host target protein binding. Our general approach enables construction of models based on other effector functions. We used SIEVE-Ub to predict nine potential novel E3 ligases from a large set of bacterial genomes. SIEVE-Ub is available for download at https://doi.org/10.6084/m9.figshare.7766984.v1 or https://github.com/biodataganache/SIEVE-Ub for the most current version.