Xuechen Mu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xuechen Mu has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Ying Xu
Systems Biology Lab for Metabolic Reprogramming, Department of Human Genetics and Cell Biology, School of Medicine, Southern University of Science and Technology, Shenzhen 518055, China
8 shared publications - Zhenyu Huang
Systems Biology Lab for Metabolic Reprogramming Department of Human Genetics and Cell Biology School of Medicine Southern University of Science and Technology Shenzhen China
7 shared publications - Qiufen Chen
Systems Biology Lab for Metabolic Reprogramming, Department of Human Genetics and Cell Biology, School of Medicine, Southern University of Science and Technology, Shenzhen 518055, China
5 shared publications - Bocheng Shi
Systems Biology Lab for Metabolic Reprogramming, Department of Human Genetics and Cell Biology, School of Medicine, Southern University of Science and Technology, Shenzhen 518055, China
4 shared publications - Kai Zhang
Department of Mathematics, Jilin University, Changchun, Jilin, China
3 shared publications - Long Xu
Systems Biology Laboratory for Metabolic Reprogramming, School of Medicine, Southern University of Science and Technology, Shenzhen 518055, China
3 shared publications
Latest Publications
- Open Access
Optimizing Model Performance and Interpretability: Application to Biological Data Classification
This study introduces a novel framework that simultaneously addresses the challenges of performance accuracy and result interpretability in transcriptomic-data-based classification. Background/objectives: In biological data classification, it is challenging to achieve both high performance accuracy and interpretability at the same time. This study presents a framework to address both challenges in transcriptomic-data-based classification. The goal is to select features, models, and a meta-voting classifier that optimizes both classification performance and interpretability. Methods: The framework consists of a four-step feature selection process: (1) the identification of metabolic pathways whose enzyme-gene expressions discriminate samples with different labels, aiding interpretability; (2) the selection of pathways whose expression variance is largely captured by the first principal component of the gene expression matrix; (3) the selection of minimal sets of genes, whose collective discerning power covers 95% of the pathway-based discerning power; and (4) the introduction of adversarial samples to identify and filter genes sensitive to such samples. Additionally, adversarial samples are used to select the optimal classification model, and a meta-voting classifier is constructed based on the optimized model results. Results: The framework applied to two cancer classification problems showed that in the binary classification, the prediction performance was comparable to the full-gene model, with F1-score differences of between −5% and 5%. In the ternary classification, the performance was significantly better, with F1-score differences ranging from −2% to 12%, while also maintaining excellent interpretability of the selected feature genes. Conclusions: This framework effectively integrates feature selection, adversarial sample handling, and model optimization, offering a valuable tool for a wide range of biological data classification problems. Its ability to balance performance accuracy and high interpretability makes it highly applicable in the field of computational biology.
Open AccessOptimizing Model Performance and Interpretability: Application to Biological Data Classification
This study introduces a novel framework that simultaneously addresses the challenges of performance accuracy and result interpretability in transcriptomic-data-based classification. Background/objectives: In biological data classification, it is challenging to achieve both high performance accuracy and interpretability at the same time. This study presents a framework to address both challenges in transcriptomic-data-based classification. The goal is to select features, models, and a meta-voting classifier that optimizes both classification performance and interpretability. Methods: The framework consists of a four-step feature selection process: (1) the identification of metabolic pathways whose enzyme-gene expressions discriminate samples with different labels, aiding interpretability; (2) the selection of pathways whose expression variance is largely captured by the first principal component of the gene expression matrix; (3) the selection of minimal sets of genes, whose collective discerning power covers 95% of the pathway-based discerning power; and (4) the introduction of adversarial samples to identify and filter genes sensitive to such samples. Additionally, adversarial samples are used to select the optimal classification model, and a meta-voting classifier is constructed based on the optimized model results. Results: The framework applied to two cancer classification problems showed that in the binary classification, the prediction performance was comparable to the full-gene model, with F1-score differences of between −5% and 5%. In the ternary classification, the performance was significantly better, with F1-score differences ranging from −2% to 12%, while also maintaining excellent interpretability of the selected feature genes. Conclusions: This framework effectively integrates feature selection, adversarial sample handling, and model optimization, offering a valuable tool for a wide range of biological data classification problems. Its ability to balance performance accuracy and high interpretability makes it highly applicable in the field of computational biology.
- Open Access
- 31 December 2024
- Published by MDPI AG in International Journal of Molecular Sciences
- Vol. 26 (1) , 275
Elucidation of Factors Affecting the Age-Dependent Cancer Occurrence Rates
Cancer occurrence rates exhibit diverse age-related patterns, and understanding them may shed new and important light on the drivers of cancer evolution. This study systematically analyzes the age-dependent occurrence rates of 23 carcinoma types, focusing on their age-dependent distribution patterns, the determinants of peak occurrence ages, and the significant difference between the two genders. According to the SEER reports, these cancer types have two types of age-dependent occurrence rate (ADOR) distributions, with most having a unimodal distribution and a few having a bimodal distribution. Our modeling analyses have revealed that (1) the first type can be naturally and simply explained using two age-dependent parameters: the total number of stem cell divisions in an organ from birth to the current age and the availability levels of bloodborne growth factors specifically needed by the cancer (sub)type, and (2) for the second type, the first peak is due to viral infection, while the second peak can be explained as in (1) for each cancer type. Further analyses indicate that (i) the iron level in an organ makes the difference between the male and female cancer occurrence rates, and (ii) the levels of sex hormones are the key determinants in the onset age of multiple cancer types. This analysis deepens our understanding of the dynamics of cancer evolution shared by diverse cancer types and provides new insights that are useful for cancer prevention and therapeutic strategies, thereby addressing critical gaps in the current paradigm of oncological research.
- 31 December 2024
- Published by MDPI AG in International Journal of Molecular Sciences
- Vol. 26 (1) , 275
Open AccessElucidation of Factors Affecting the Age-Dependent Cancer Occurrence Rates
Cancer occurrence rates exhibit diverse age-related patterns, and understanding them may shed new and important light on the drivers of cancer evolution. This study systematically analyzes the age-dependent occurrence rates of 23 carcinoma types, focusing on their age-dependent distribution patterns, the determinants of peak occurrence ages, and the significant difference between the two genders. According to the SEER reports, these cancer types have two types of age-dependent occurrence rate (ADOR) distributions, with most having a unimodal distribution and a few having a bimodal distribution. Our modeling analyses have revealed that (1) the first type can be naturally and simply explained using two age-dependent parameters: the total number of stem cell divisions in an organ from birth to the current age and the availability levels of bloodborne growth factors specifically needed by the cancer (sub)type, and (2) for the second type, the first peak is due to viral infection, while the second peak can be explained as in (1) for each cancer type. Further analyses indicate that (i) the iron level in an organ makes the difference between the male and female cancer occurrence rates, and (ii) the levels of sex hormones are the key determinants in the onset age of multiple cancer types. This analysis deepens our understanding of the dynamics of cancer evolution shared by diverse cancer types and provides new insights that are useful for cancer prevention and therapeutic strategies, thereby addressing critical gaps in the current paradigm of oncological research.
- Open Access
- 2 December 2024
- Published by MDPI AG in International Journal of Molecular Sciences
- Vol. 25 (23) , 12942
DeepEnhancerPPO: An Interpretable Deep Learning Approach for Enhancer Classification
Enhancers are short genomic segments located in non-coding regions of the genome that play a critical role in regulating the expression of target genes. Despite their importance in transcriptional regulation, effective methods for classifying enhancer categories and regulatory strengths remain limited. To address this challenge, we propose a novel end-to-end deep learning architecture named DeepEnhancerPPO. The model integrates ResNet and Transformer modules to extract local, hierarchical, and long-range contextual features. Following feature fusion, we employ Proximal Policy Optimization (PPO), a reinforcement learning technique, to reduce the dimensionality of the fused features, retaining the most relevant features for downstream classification tasks. We evaluate the performance of DeepEnhancerPPO from multiple perspectives, including ablation analysis, independent tests, assessment of PPO’s contribution to performance enhancement, and interpretability of the classification results. Each module positively contributes to the overall performance, with ResNet and PPO being the most significant contributors. Overall, DeepEnhancerPPO demonstrates superior performance on independent datasets compared to other models, outperforming the second-best model by 6.7% in accuracy for enhancer category classification. The model consistently ranks among the top five classifiers out of 25 for enhancer strength classification without requiring re-optimization of the hyperparameters and ranks as the second-best when the hyperparameters are refined. This indicates that the DeepEnhancerPPO framework is highly robust for enhancer classification. Additionally, the incorporation of PPO enhances the interpretability of the classification results.
- 2 December 2024
- Published by MDPI AG in International Journal of Molecular Sciences
- Vol. 25 (23) , 12942
Open AccessDeepEnhancerPPO: An Interpretable Deep Learning Approach for Enhancer Classification
Enhancers are short genomic segments located in non-coding regions of the genome that play a critical role in regulating the expression of target genes. Despite their importance in transcriptional regulation, effective methods for classifying enhancer categories and regulatory strengths remain limited. To address this challenge, we propose a novel end-to-end deep learning architecture named DeepEnhancerPPO. The model integrates ResNet and Transformer modules to extract local, hierarchical, and long-range contextual features. Following feature fusion, we employ Proximal Policy Optimization (PPO), a reinforcement learning technique, to reduce the dimensionality of the fused features, retaining the most relevant features for downstream classification tasks. We evaluate the performance of DeepEnhancerPPO from multiple perspectives, including ablation analysis, independent tests, assessment of PPO’s contribution to performance enhancement, and interpretability of the classification results. Each module positively contributes to the overall performance, with ResNet and PPO being the most significant contributors. Overall, DeepEnhancerPPO demonstrates superior performance on independent datasets compared to other models, outperforming the second-best model by 6.7% in accuracy for enhancer category classification. The model consistently ranks among the top five classifiers out of 25 for enhancer strength classification without requiring re-optimization of the hyperparameters and ranks as the second-best when the hyperparameters are refined. This indicates that the DeepEnhancerPPO framework is highly robust for enhancer classification. Additionally, the incorporation of PPO enhances the interpretability of the classification results.
- Open Access
- 25 November 2024
- Published by MDPI AG in Brain Sciences
- Vol. 14 (12) , 1180
Construction of a Dataset for All Expressed Transcripts for Alzheimer’s Disease Research
Accurate identification and functional annotation of splicing isoforms and non-coding RNAs (lncRNAs), alongside full-length protein-encoding transcripts, are critical for understanding gene (mis)regulation and metabolic reprogramming in Alzheimer’s disease (AD). This study aims to provide a comprehensive and accurate transcriptome resource to improve existing AD transcript databases. Background/Objectives: Gene mis-regulation and metabolic reprogramming play a key role in AD, yet existing transcript databases lack accurate and comprehensive identification of splicing isoforms and lncRNAs. This study aims to generate a refined transcriptome dataset, expanding the understanding of AD onset and progression. Methods: Publicly available RNA-seq data from pre-AD and AD tissues were utilized. Advanced bioinformatics tools were applied to assemble and annotate full-length transcripts, including splicing isoforms and lncRNAs, with an emphasis on correcting errors and enhancing annotation accuracy. Results: A significantly improved transcriptome dataset was generated, which includes detailed annotations of splicing isoforms and lncRNAs. This dataset expands the scope of existing AD transcript databases and provides new insights into the molecular mechanisms underlying AD. The findings demonstrate that the refined dataset captures more relevant details about AD progression compared to publicly available data. Conclusions: The newly developed transcriptome resource and the associated analysis tools offer a valuable contribution to AD research, providing deeper insights into the disease’s molecular mechanisms. This work supports future research into gene regulation and metabolic reprogramming in AD and serves as a foundation for exploring novel therapeutic targets.
- 25 November 2024
- Published by MDPI AG in Brain Sciences
- Vol. 14 (12) , 1180
Open AccessConstruction of a Dataset for All Expressed Transcripts for Alzheimer’s Disease Research
Accurate identification and functional annotation of splicing isoforms and non-coding RNAs (lncRNAs), alongside full-length protein-encoding transcripts, are critical for understanding gene (mis)regulation and metabolic reprogramming in Alzheimer’s disease (AD). This study aims to provide a comprehensive and accurate transcriptome resource to improve existing AD transcript databases. Background/Objectives: Gene mis-regulation and metabolic reprogramming play a key role in AD, yet existing transcript databases lack accurate and comprehensive identification of splicing isoforms and lncRNAs. This study aims to generate a refined transcriptome dataset, expanding the understanding of AD onset and progression. Methods: Publicly available RNA-seq data from pre-AD and AD tissues were utilized. Advanced bioinformatics tools were applied to assemble and annotate full-length transcripts, including splicing isoforms and lncRNAs, with an emphasis on correcting errors and enhancing annotation accuracy. Results: A significantly improved transcriptome dataset was generated, which includes detailed annotations of splicing isoforms and lncRNAs. This dataset expands the scope of existing AD transcript databases and provides new insights into the molecular mechanisms underlying AD. The findings demonstrate that the refined dataset captures more relevant details about AD progression compared to publicly available data. Conclusions: The newly developed transcriptome resource and the associated analysis tools offer a valuable contribution to AD research, providing deeper insights into the disease’s molecular mechanisms. This work supports future research into gene regulation and metabolic reprogramming in AD and serves as a foundation for exploring novel therapeutic targets.
- Open Access
- 1 October 2024
- Published by Springer Nature in Research Square
DeepEnhancerPPO: An Interpretable Deep Learning Approach for Enhancer Classification
Enhancers are short genomic segments located in non-coding regions in a genome that help to increase the expressions of the target genes. Despite their significance in transcription regulation, effective methods for classifying enhancer categories and regulatory strengths remain limited. To address the issue, we propose a novel end-to-end deep learning architecture named DeepEnhancerPPO. The model integrates ResNet and Transformer modules to extract local, hierarchical, and long-range contextual features. Following feature fusion, we employ the proximal policy optimization (PPO), a reinforcement learning technique, to reduce the dimensionality of the fused features, retaining the most relevant ones for downstream classification. We evaluate the performance of DeepEnhancerPPO from multiple perspectives, including ablation analysis, independent tests, and interpretability of classification results. Each of these modules contributes positively to the model's performance, with ResNet and PPO being the top contributors. Overall, DeepEnhancerPPO exhibits superb performance on independent datasets compared to other models, outperforming the second-best model by 6.7% in accuracy for enhancer category classification. The model also ranks within the top five classifiers out of 25 in enhancer strength classification without the need to re-optimize the hyperparameters, indicating that the DeepEnhancerPPO framework is highly robust for enhancer classification. Additionally, the inclusion of PPO enhances the interpretability of the classification results. The source code is openly accessible at https://github.com/Mxc666/DeepEnhancerPPO.git.
- 1 October 2024
- Published by Springer Nature in Research Square
Open AccessDeepEnhancerPPO: An Interpretable Deep Learning Approach for Enhancer Classification
Enhancers are short genomic segments located in non-coding regions in a genome that help to increase the expressions of the target genes. Despite their significance in transcription regulation, effective methods for classifying enhancer categories and regulatory strengths remain limited. To address the issue, we propose a novel end-to-end deep learning architecture named DeepEnhancerPPO. The model integrates ResNet and Transformer modules to extract local, hierarchical, and long-range contextual features. Following feature fusion, we employ the proximal policy optimization (PPO), a reinforcement learning technique, to reduce the dimensionality of the fused features, retaining the most relevant ones for downstream classification. We evaluate the performance of DeepEnhancerPPO from multiple perspectives, including ablation analysis, independent tests, and interpretability of classification results. Each of these modules contributes positively to the model's performance, with ResNet and PPO being the top contributors. Overall, DeepEnhancerPPO exhibits superb performance on independent datasets compared to other models, outperforming the second-best model by 6.7% in accuracy for enhancer category classification. The model also ranks within the top five classifiers out of 25 in enhancer strength classification without the need to re-optimize the hyperparameters, indicating that the DeepEnhancerPPO framework is highly robust for enhancer classification. Additionally, the inclusion of PPO enhances the interpretability of the classification results. The source code is openly accessible at https://github.com/Mxc666/DeepEnhancerPPO.git.
- Open Access
- 25 August 2024
- Published by MDPI AG in International Journal of Molecular Sciences
- Vol. 25 (17) , 9211
Elucidating the Functional Roles of Long Non-Coding RNAs in Alzheimer’s Disease
Alzheimer’s disease (AD) is a multifaceted neurodegenerative disorder characterized by cognitive decline and neuronal loss, representing a most challenging health issue. We present a computational analysis of transcriptomic data of AD tissues vs. healthy controls, focused on the elucidation of functional roles played by long non-coding RNAs (lncRNAs) throughout the AD progression. We first assembled our own lncRNA transcripts from the raw RNA-Seq data generated from 527 samples of the dorsolateral prefrontal cortex, resulting in the identification of 31,574 novel lncRNA genes. Based on co-expression analyses between mRNAs and lncRNAs, a co-expression network was constructed. Maximal subnetworks with dense connections were identified as functional clusters. Pathway enrichment analyses were conducted over mRNAs and lncRNAs in each cluster, which served as the basis for the inference of functional roles played by lncRNAs involved in each of the key steps in an AD development model that we have previously built based on transcriptomic data of protein-encoding genes. Detailed information is presented about the functional roles of lncRNAs in activities related to stress response, reprogrammed metabolism, cell polarity, and development. Our analyses also revealed that lncRNAs have the discerning power to distinguish between AD samples of each stage and healthy controls. This study represents the first of its kind.
- 25 August 2024
- Published by MDPI AG in International Journal of Molecular Sciences
- Vol. 25 (17) , 9211
Open AccessElucidating the Functional Roles of Long Non-Coding RNAs in Alzheimer’s Disease
Alzheimer’s disease (AD) is a multifaceted neurodegenerative disorder characterized by cognitive decline and neuronal loss, representing a most challenging health issue. We present a computational analysis of transcriptomic data of AD tissues vs. healthy controls, focused on the elucidation of functional roles played by long non-coding RNAs (lncRNAs) throughout the AD progression. We first assembled our own lncRNA transcripts from the raw RNA-Seq data generated from 527 samples of the dorsolateral prefrontal cortex, resulting in the identification of 31,574 novel lncRNA genes. Based on co-expression analyses between mRNAs and lncRNAs, a co-expression network was constructed. Maximal subnetworks with dense connections were identified as functional clusters. Pathway enrichment analyses were conducted over mRNAs and lncRNAs in each cluster, which served as the basis for the inference of functional roles played by lncRNAs involved in each of the key steps in an AD development model that we have previously built based on transcriptomic data of protein-encoding genes. Detailed information is presented about the functional roles of lncRNAs in activities related to stress response, reprogrammed metabolism, cell polarity, and development. Our analyses also revealed that lncRNAs have the discerning power to distinguish between AD samples of each stage and healthy controls. This study represents the first of its kind.
- Open Access
- 22 July 2024
- Published by Springer Nature in Research Square
Optimizing Model Performance and Interpretability: an application to biological data classification
In biological data classification, both performance accuracy and result interpretability are desired and yet difficult to achieve simultaneously. We present a framework for transcriptomic data-based classification that can accomplish both. The key idea is as follows: 1) to identify metabolic pathways whose expressions have strong discerning power in separating samples having distinct labels, hence providing a basis for providing interpretability of the classification results; 2) to select pathways from the afore-identified whose expression variance for each can be largely captured by its first principal component of the gene-expression matrix for the pathway, hence allowing to select a minimal number of discerning pathways; 3) to select a minimal set of genes whose collective discerning power covers 95% of the discerning power for each selected pathway, giving rise to a set of features (genes) for classification; and 4) to select a model among the available ones and model parameters that give the optimal classification results. We have demonstrated the effectiveness of this framework on two cancer biology problems. We anticipate that this framework will be used for the selection of features, model, and model parameters for a wide range of biological data classification problems.
- 22 July 2024
- Published by Springer Nature in Research Square
Open AccessOptimizing Model Performance and Interpretability: an application to biological data classification
In biological data classification, both performance accuracy and result interpretability are desired and yet difficult to achieve simultaneously. We present a framework for transcriptomic data-based classification that can accomplish both. The key idea is as follows: 1) to identify metabolic pathways whose expressions have strong discerning power in separating samples having distinct labels, hence providing a basis for providing interpretability of the classification results; 2) to select pathways from the afore-identified whose expression variance for each can be largely captured by its first principal component of the gene-expression matrix for the pathway, hence allowing to select a minimal number of discerning pathways; 3) to select a minimal set of genes whose collective discerning power covers 95% of the discerning power for each selected pathway, giving rise to a set of features (genes) for classification; and 4) to select a model among the available ones and model parameters that give the optimal classification results. We have demonstrated the effectiveness of this framework on two cancer biology problems. We anticipate that this framework will be used for the selection of features, model, and model parameters for a wide range of biological data classification problems.
- Open Access
- 10 July 2024
- Published by MDPI AG in Preprints.org
Elucidating the Functional Roles of Long Non-coding RNAs in Alzheimer's Disease
Alzheimer's disease (AD) is a multifaceted neurodegenerative disorder characterized by cognitive decline and neuronal loss, representing a most challenging health issue. We present a computational analysis of transcriptomic data of AD tissues vs. healthy controls, focused on elucidation of functional roles played by long non-coding RNAs (lncRNAs) throughout the AD progression. We first assembled our own lncRNA transcripts from the raw RNA-Seq data generated from 527 samples of dorsolateral prefrontal cortex, resulting in the identification of 31,574 novel lncRNA genes. Based on co-expression analyses between mRNAs and lncRNAs, a co-expression network is constructed. Maximal subnetworks with dense connections are identified as functional clusters. Pathway enrichment analyses are conducted over mRNAs and lncRNAs in each cluster, which serve as the basis for the inference of functional roles played by lncRNAs involved in each of the key steps in an AD development model that we have previously build based on transcriptomic data of protein-encoding genes. Detailed information is presented about the functional roles by lncRNAs in activities related to stress response, reprogrammed metabolism, cell-polarity, and development. Our analyses have also revealed that lncRNAs have discerning power in distinguishing between AD samples of each stage and healthy controls. This study represents the first of its kind.
- 10 July 2024
- Published by MDPI AG in Preprints.org
Open AccessElucidating the Functional Roles of Long Non-coding RNAs in Alzheimer's Disease
Alzheimer's disease (AD) is a multifaceted neurodegenerative disorder characterized by cognitive decline and neuronal loss, representing a most challenging health issue. We present a computational analysis of transcriptomic data of AD tissues vs. healthy controls, focused on elucidation of functional roles played by long non-coding RNAs (lncRNAs) throughout the AD progression. We first assembled our own lncRNA transcripts from the raw RNA-Seq data generated from 527 samples of dorsolateral prefrontal cortex, resulting in the identification of 31,574 novel lncRNA genes. Based on co-expression analyses between mRNAs and lncRNAs, a co-expression network is constructed. Maximal subnetworks with dense connections are identified as functional clusters. Pathway enrichment analyses are conducted over mRNAs and lncRNAs in each cluster, which serve as the basis for the inference of functional roles played by lncRNAs involved in each of the key steps in an AD development model that we have previously build based on transcriptomic data of protein-encoding genes. Detailed information is presented about the functional roles by lncRNAs in activities related to stress response, reprogrammed metabolism, cell-polarity, and development. Our analyses have also revealed that lncRNAs have discerning power in distinguishing between AD samples of each stage and healthy controls. This study represents the first of its kind.
- Open Access
- 17 March 2021
- Published by MDPI AG in International Journal of Molecular Sciences
- Vol. 22 (6) , 3079
A Novel Position-Specific Encoding Algorithm (SeqPose) of Nucleotide Sequences and Its Application for Detecting Enhancers
Enhancers are short genomic regions exerting tissue-specific regulatory roles, usually for remote coding regions. Enhancers are observed in both prokaryotic and eukaryotic genomes, and their detections facilitate a better understanding of the transcriptional regulation mechanism. The accurate detection and transcriptional regulation strength evaluation of the enhancers remain a major bioinformatics challenge. Most of the current studies utilized the statistical features of short fixed-length nucleotide sequences. This study introduces the location information of each k-mer (SeqPose) into the encoding strategy of a DNA sequence and employs the attention mechanism in the two-layer bi-directional long-short term memory (BD-LSTM) model (spEnhancer) for the enhancer detection problem. The first layer of the delivered classifier discriminates between enhancers and non-enhancers, and the second layer evaluates the transcriptional regulation strength of the detected enhancer. The SeqPose-encoded features are selected by the Chi-squared test, and 45 positions are removed from further analysis. The existing studies may focus on selecting the statistical DNA sequence descriptors with large contributions to the prediction models. This study does not utilize these statistical DNA sequence descriptors. Then the word vector of the SeqPose-encoded features is obtained by using the word embedding layer. This study hypothesizes that different word vector features may contribute differently to the enhancer detection model, and assigns different weights to these word vectors through the attention mechanism in the BD-LSTM model. The previous study generously provided the training and independent test datasets, and the proposed spEnhancer is compared with the three existing state-of-the-art studies using the same experimental procedure. The leave-one-out validation data on the training dataset shows that the proposed spEnhancer achieves similar detection performances as the three existing studies. While spEnhancer achieves the best overall performance metric MCC for both of the two binary classification problems on the independent test dataset. The experimental data shows that the strategy of removing redundant positions (SeqPose) may help improve the DNA sequence-based prediction models. spEnhancer may serve well as a complementary model to the existing studies, especially for the novel query enhancers that are not included in the training dataset.
- 17 March 2021
- Published by MDPI AG in International Journal of Molecular Sciences
- Vol. 22 (6) , 3079
Open AccessA Novel Position-Specific Encoding Algorithm (SeqPose) of Nucleotide Sequences and Its Application for Detecting Enhancers
Enhancers are short genomic regions exerting tissue-specific regulatory roles, usually for remote coding regions. Enhancers are observed in both prokaryotic and eukaryotic genomes, and their detections facilitate a better understanding of the transcriptional regulation mechanism. The accurate detection and transcriptional regulation strength evaluation of the enhancers remain a major bioinformatics challenge. Most of the current studies utilized the statistical features of short fixed-length nucleotide sequences. This study introduces the location information of each k-mer (SeqPose) into the encoding strategy of a DNA sequence and employs the attention mechanism in the two-layer bi-directional long-short term memory (BD-LSTM) model (spEnhancer) for the enhancer detection problem. The first layer of the delivered classifier discriminates between enhancers and non-enhancers, and the second layer evaluates the transcriptional regulation strength of the detected enhancer. The SeqPose-encoded features are selected by the Chi-squared test, and 45 positions are removed from further analysis. The existing studies may focus on selecting the statistical DNA sequence descriptors with large contributions to the prediction models. This study does not utilize these statistical DNA sequence descriptors. Then the word vector of the SeqPose-encoded features is obtained by using the word embedding layer. This study hypothesizes that different word vector features may contribute differently to the enhancer detection model, and assigns different weights to these word vectors through the attention mechanism in the BD-LSTM model. The previous study generously provided the training and independent test datasets, and the proposed spEnhancer is compared with the three existing state-of-the-art studies using the same experimental procedure. The leave-one-out validation data on the training dataset shows that the proposed spEnhancer achieves similar detection performances as the three existing studies. While spEnhancer achieves the best overall performance metric MCC for both of the two binary classification problems on the independent test dataset. The experimental data shows that the strategy of removing redundant positions (SeqPose) may help improve the DNA sequence-based prediction models. spEnhancer may serve well as a complementary model to the existing studies, especially for the novel query enhancers that are not included in the training dataset.