Dataset Information

Handling class imbalance problem in miRNA dataset associated with cancer.

ABSTRACT: MiRNAs are small (~22nt long) non-coding RNA sequences; binds to the complementarity target sites in 3' Untranslated Region (UTR) of mRNA sequences but not restricted to other mRNA regions viz., 5' UTR and Coding sequences (CDS). Complementarity binding of miRNA to mRNA target sites either results in complete degradation of the mRNA itself or it may regulate the mRNA as an oncogene or as a tumor suppressor gene. However, the exact mechanism involved in identifying a miRNA to be associated with cancer is still unclear. Further, with the outburst in the number of miRNAs sequences recorded every year in miRBase, the gap is still widening mainly due to the laborious and economically unfavorable experimental procedures associated with the functional annotation. Motivated by the fact, we constructed a two-step support vector machine-based predictive model - miRSEQ and miRINT. However, the major pitfall during the construction of the model is the class imbalance problem. Hence, in order to overcome class imbalance problem, in the present study we empirically compare the effectiveness of two different methods viz., Synthetic Minority Oversampling Technique (SMOTE) and cost-senstive learning method. Performance measures were evaluated in terms of Precision and Recall. Based on our result, it was observed that for miRNA dataset with high class imbalance utilized for predicting association of cancer, cost-sensitive method outperformed the oversampling method.

SUBMITTER: Kothandan R

PROVIDER: S-EPMC4349932 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundModeling patient data, particularly electronic health records (EHR), is one of the major focuses of machine learning studies in healthcare, as these records provide clinicians with valuable information that can potentially assist them in disease diagnosis and decision-making.MethodsIn this study, we present a multi-level graph-based framework called MedMGF, which models both patient medical profiles extracted from EHR data and their relationship network of health profiles in a single architecture. The medical profiles consist of several layers of data embedding derived from interval records obtained during hospitalization, and the patient-patient network is created by measuring the similarities between these profiles. We also propose a modification to the Focal Loss (FL) function to improve classification performance in imbalanced datasets without the need to imputate the data. MedMGF's performance was evaluated against several Graphical Convolutional Network (GCN) baseline models implemented with Binary Cross Entropy (BCE), FL, class balancing parameter α , and Synthetic Minority Oversampling Technique (SMOTE).ResultsOur proposed framework achieved high classification performance (AUC: 0.8098, ACC: 0.7503, SEN: 0.8750, SPE: 0.7445, NPV: 0.9923, PPV: 0.1367) on an extreme imbalanced pediatric sepsis dataset (n=3,014, imbalance ratio of 0.047). It yielded a classification improvement of 3.81% for AUC, 15% for SEN compared to the baseline GCN+ α FL (AUC: 0.7717, ACC: 0.8144, SEN: 0.7250, SPE: 0.8185, PPV: 0.1559, NPV: 0.9847), and an improvement of 5.88% in AUC and 22.5% compared to GCN+FL+SMOTE (AUC: 0.7510, ACC: 0.8431, SEN: 0.6500, SPE: 0.8520, PPV: 0.1688, NPV: 0.9814). It also showed a classification improvement of 3.86% for AUC, 15% for SEN compared to the baseline GCN+ α BCE (AUC: 0.7712, ACC: 0.8133, SEN: 0.7250, SPE: 0.8173, PPV: 0.1551, NPV: 0.9847), and an improvement of 14.33% in AUC and 27.5% in comparison to GCN+BCE+SMOTE (AUC: 0.6665, ACC: 0.7271, SEN: 0.6000, SPE: 0.7329, PPV: 0.0941, NPV: 0.9754).ConclusionWhen compared to all baseline models, MedMGF achieved the highest SEN and AUC results, demonstrating the potential for several healthcare applications.

Dataset Information

Handling class imbalance problem in miRNA dataset associated with cancer.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets