DeepRF: A deep learning method for predicting metabolic pathways in organisms based on annotated genomes

Shah, Hayat Ali; Liu, Juan; Yang, Zhihui; Zhang, Xiaolei; Feng, Jing

doi:10.1016/j.compbiomed.2022.105756

Cited by 6 publications

(7 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the representative machine learning-based pathway prediction methods we included both results from the mlLGPR and triUMPF [ 17 ] studies. The models from Aljarbou et al [ 12 ] and DeepRF [ 13 ] were not used in the evaluation because both models are binary classifiers instead of multi-label and are trained using different datasets making it difficult to accurately compare. In addition, from the best of our knowledge the datasets and source code used in both studies are not open source which make comparing their performances even more difficult.…”

Section: Methodsmentioning

confidence: 99%

“…Despite the promising results from the study, PathoLogic is still used as the main engine for Pathway Tool’s prediction algorithm. Recently, there has been several studies which updated the pioneer study with new datasets, features and methodologies [ 11 – 13 ].…”

Section: Introductionmentioning

confidence: 99%

“…Modeling the prediction task as multi-label classification allowed the training dataset used in mlLGPR to be more compact allowing for more organisms to be used for training. For example, Aljarbou et al [ 12 ] has 4979 instances covering 20 organisms and DeepRF [ 13 ] had 172,380 instances covering 60 organisms. mlLGPR’s multi-label modeling allows for its dataset to be smaller with 15,000 instances but is able to cover 15,000 organisms.…”

Section: Introductionmentioning

confidence: 99%

“…mlLGPR uses a binary relevance approach [ 15 ] where the multi-label learning process is divided into independent binary classifiers for each pathway label allowing for the possibility of parallel training. What also differentiated mlLGPR with other pathway prediction studies such as [ 1 , 12 , 13 ] is that for their evaluation methodology they used a completely separate evaluation dataset which they did not use for training and hyperparameter tuning. Another novel contribution from the mlLGPR study was that it was the first machine learning-based pathway prediction method to be evaluated also on multi-organism genomes such as symbionts and microbiomes.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Multi-label classification with XGBoost for metabolic pathway prediction

Joe,

Kim

2024

BMC Bioinformatics

View full text Add to dashboard Cite

Background Metabolic pathway prediction is one possible approach to address the problem in system biology of reconstructing an organism’s metabolic network from its genome sequence. Recently there have been developments in machine learning-based pathway prediction methods that conclude that machine learning-based approaches are similar in performance to the most used method, PathoLogic which is a rule-based method. One issue is that previous studies evaluated PathoLogic without taxonomic pruning which decreases its performance. Results In this study, we update the evaluation results from previous studies to demonstrate that PathoLogic with taxonomic pruning outperforms previous machine learning-based approaches and that further improvements in performance need to be made for them to be competitive. Furthermore, we introduce mlXGPR, a XGBoost-based metabolic pathway prediction method based on the multi-label classification pathway prediction framework introduced from mlLGPR. We also improve on this multi-label framework by utilizing correlations between labels using classifier chains. We propose a ranking method that determines the order of the chain so that lower performing classifiers are placed later in the chain to utilize the correlations between labels more. We evaluate mlXGPR with and without classifier chains on single-organism and multi-organism benchmarks. Our results indicate that mlXGPR outperform other previous pathway prediction methods including PathoLogic with taxonomic pruning in terms of hamming loss, precision and F1 score on single organism benchmarks. Conclusions The results from our study indicate that the performance of machine learning-based pathway prediction methods can be substantially improved and can even outperform PathoLogic with taxonomic pruning.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multi-label classification with XGBoost for metabolic pathway prediction

Joe,

Kim

2024

BMC Bioinformatics

View full text Add to dashboard Cite

show abstract

“…Recent methods utilizing learning-and network-based algorithms are on the rise to overcome these challenges and decode causal relations between omic entities [8][9][10][11] . Learning based methods efficiently integrate multi-omic data to extract interpretable annotations such as pathways, reactions, and processes [12][13][14] . Also, network-based algorithms, including shortest paths 15 , Steiner trees/forests 16,17 , and random walk 18,19 have been frequently used to construct specific networks by propagating omic hits 20,21 .…”

Section: Introductionmentioning

confidence: 99%

Unveiling Hidden Connections in Omics DataviapyPARAGON: an Integrative Hybrid Approach for Disease Network Construction

Arici,

Tuncbag

2023

Preprint

View full text Add to dashboard Cite

Omics technologies are powerful tools for detecting dysregulated and altered signaling components in various contexts, encompassing disease states, patients, and drug-perturbations. Network inference or reconstruction algorithms play an integral role in the successful analysis and identification of causal relationships between omics hits. However, accurate representation of signaling networks and identification of context-specific interactions within sparse omics datasets in complex interactomes pose significant challenges in integrative approaches. To address these challenges, we present pyPARAGON (PAgeRAnk-flux on Graphlet-guided network for multi-Omic data integratioN), a novel tool that combines network propagation with graphlets. By leveraging network motifs instead of pairwise connections among proteins, pyPARAGON offers improved accuracy and reduces the inclusion of nonspecific interactions in signaling networks. Through comprehensive evaluations on benchmark cancer signaling pathways, we demonstrate that pyPARAGON outperforms state-of-the-art approaches in node propagation and edge inference. Furthermore, pyPARAGON exhibits promising performance in discovering cancer driver networks. Notably, we demonstrate its utility in network-based stratification of patient tumors by integrating phosphoproteomic data from 105 breast cancer tumors with the interactome, leading to the discovery of tumor-specific signaling pathways. Overall, the development and evaluation of pyPARAGON significantly contributes to the field as an effective tool for the analysis and integration of multi-omic data in the context of signaling networks. pyPARAGON is available at https://github.com/metunetlab/pyPARAGON.

show abstract

Artificial intelligence in systems biology

Dasgupta

2023

Handbook of Statistics

View full text Add to dashboard Cite

DeepRF: A deep learning method for predicting metabolic pathways in organisms based on annotated genomes

Cited by 6 publications

References 22 publications

Multi-label classification with XGBoost for metabolic pathway prediction

Multi-label classification with XGBoost for metabolic pathway prediction

Unveiling Hidden Connections in Omics DataviapyPARAGON: an Integrative Hybrid Approach for Disease Network Construction

Artificial intelligence in systems biology

Contact Info

Product

Resources

About