Human Protein Subcellular Localization with Integrated Source and Multi-label Ensemble Classifier

Guo, Xiaotong; Liu, Fulin; Ju, Ying; Wang, Zhen; Wang, Chunyu

doi:10.1038/srep28087

Cited by 43 publications

(34 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The experimental result denotes that our proposed method works better than all the traditional methods used in [13]. Our proposed method achieves 69.1% average precision, almost 15% higher than MLKNN, which is the best result in [13]. Our method achieves higher Micro-averaged F-Measure score than traditional methods, about 16% improvement.…”

Section: ) Rank-based Evaluation Metricsmentioning

confidence: 79%

“…The higher value, the better performance: Table 1 shows the comparison result of our model with several traditional methods, such as MLKNN and BRKNN. The experimental result denotes that our proposed method works better than all the traditional methods used in [13]. Our proposed method achieves 69.1% average precision, almost 15% higher than MLKNN, which is the best result in [13].…”

Section: ) Rank-based Evaluation Metricsmentioning

confidence: 81%

“…The dataset used in this work are obtained from [13].There are 4802 different protein sequences, among which 3448 are single-location proteins, 1354 are multiple-location proteins. The protein sequences used in this work are provided as a set of flat files in fasta format, each sequence is represented as an identifier and lines of data.…”

Section: A Datasetmentioning

confidence: 99%

See 2 more Smart Citations

A Multi-label Classifier for Human Protein Subcellular Localization Based on LSTM Networks

Zhi-ying¹,

Sun²,

Wei³

2018

Proceedings of the 2018 International Conference on Advanced Control, Automation and Artificial Intelligence (ACAAI 2018)

View full text Add to dashboard Cite

Nowadays, with the increasing number of protein sequences all over the world, more and more people are paying their attention to predicting protein subcellular location. Since wet experiment is costly and time-consuming, the automatic computational methods are urgent. In this paper, we propose a variant model based on Long Short-Term Memory (LSTM) to predict protein subcellular location. In this model, we employ LSTM to capture long distance dependency features of the sequence data. Moreover, we adjust the loss function of the loss layer to solve multi-label classification problem. Experimental results demonstrate that, compared with the traditional machine learning methods, our method achieves the best performance in various evaluation metrics.

show abstract

Section: ) Rank-based Evaluation Metricsmentioning

confidence: 79%

Section: ) Rank-based Evaluation Metricsmentioning

confidence: 81%

See 1 more Smart Citation

A Multi-label Classifier for Human Protein Subcellular Localization Based on LSTM Networks

Zhi-ying¹,

Sun²,

Wei³

2018

Proceedings of the 2018 International Conference on Advanced Control, Automation and Artificial Intelligence (ACAAI 2018)

View full text Add to dashboard Cite

show abstract

“…, MRF − l t M . This process is repeated iteratively until the convergence of the pseudo-likelihood Equation 10. We name these MRFs as collective MRFs (see Figure 2 and algorithm 1).…”

Section: Collective Mrfsmentioning

confidence: 99%

“…Protein features, especially the sequence-based features, are always the essential part in various protein SCL predictors [4,7,10]. To carry out di erent functions, one protein can be located in di erent SubCellular Compartments (SCCs) simultaneously or at di erent times during di erent biological processes, e.g.…”

Section: Introductionmentioning

confidence: 99%

Bayesian Collective Markov Random Fields for Subcellular Localization Prediction of Human Proteins

Zhu

Ester

2017

Proceedings of the 8th ACM International Conference on Bioinformatics, Computational Biology,and Health Informatics

View full text Add to dashboard Cite

Advanced biotechnology makes it possible to access a multitude of heterogeneous proteomic, interactomic, genomic, and functional annotation data. One challenge in computational biology is to integrate these data to enable automated prediction of the Subcellular Localizations (SCL) of human proteins. For proteins that have multiple biological roles, their correct in silico assignment to di erent SCL can be considered as an imbalanced multi-label classi cation problem. In this study, we developed a Bayesian Collective Markov Random Fields (BCMRFs) model for multi-SCL prediction of human proteins. Given a set of unknown proteins and their corresponding protein-protein interaction (PPI) network, the SCLs of each protein can be inferred by the SCLs of its interacting partners. To do so, we integrate PPIs, the adjacency of SCLs and protein features, and perform transductive learning on the re-balanced dataset. Our experimental results show that the spatial adjacency of the SCLs improves multi-SCL prediction, especially for the SCLs with few annotated instances. Our approach outperforms the state-of-art PPIbased and feature-based multi-SCL prediction method for human proteins. KEYWORDSHuman protein subcellular localization; markov random eld; transductive learning; imbalanced multi-label classi cation.

show abstract

HPSLPred: An Ensemble Multi‐Label Classifier for Human Protein Subcellular Location Prediction with Imbalanced Source

2017

View full text Add to dashboard Cite

Protein subcellular localization prediction is an important and challenging problem. The traditional biology experiments are expensive and time-consuming, so more and more research interests tend to a series of machine learning approaches for predicting protein subcellular location. There are two main difficult problems among the existing state-of-the-art methods. First, most of the existing techniques are designed to deal with the multi-class but not the multi-label classification, which ignores the connection between the multiple labels. In reality, multiple location proteins implicate that there are vital and unique biological significances worthy of special focus, which cannot be ignored. Second, the techniques for handling imbalanced data in multi-label classification problem is significant but less. For solving the two issues, we have developed an ensemble multi-label classifier called HPSLPred which can be applied for the multi-label classification with imbalanced protein source. For the conveniences of users, a user-friendly webserver for HPSLPred was established at http://server.malab.cn/HPSLPred.

show abstract

Human Protein Subcellular Localization with Integrated Source and Multi-label Ensemble Classifier

Cited by 43 publications

References 55 publications

A Multi-label Classifier for Human Protein Subcellular Localization Based on LSTM Networks

A Multi-label Classifier for Human Protein Subcellular Localization Based on LSTM Networks

Bayesian Collective Markov Random Fields for Subcellular Localization Prediction of Human Proteins

HPSLPred: An Ensemble Multi‐Label Classifier for Human Protein Subcellular Location Prediction with Imbalanced Source

Contact Info

Product

Resources

About