A comparison of single imputation and multiple imputation methods for missing data in different oncogene expression profiles

Abstract

To evaluate the effects of multiple-imputation (MI) method for missing data in gene expression profiles with different datasets and percentages of missing values compared with 3 single-imputation (SI) methods. Based on 3 gene expression profiles datasets from human colon cancer, non-small cell lung cancer, and lymph cancer, different deletion rates and different imputation numbers of MI were compared. The imputation and clustering effects of different methods were evaluated using the NRMSE and the gene clustering accuracy (F value). The NRMSE of the 4 methods gradually increased as the percentage of missing values in the 3 datasets increased, whereas the F value gradually decreased. In all datasets with different percentage of missing values settings, the NRMSEs of MI was consistently lower than those of the 3 SI methods, whereas the F value of MI was highest. The NRMSEs of MI gradually decreased as the number of imputations increased and increased as the variability in the original datasets increased, and the datasets imputed by MI showed the best clustering results. The results showed that the application of MI develops and enriches imputation-model approaches and provides a solid foundation for subsequent establishment of imputation strategies for gene expression profiles with missing data.

KEYWORDS:

• Gene expression profiles
• missing data
• multiple imputation
• clustering analysis

Acknowledgments

We are grateful to the Editor, Associate Editor, and two referees for their constructive comments on the article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by National Natural Science Foundation of China: [Grant Number 81273178, 81872716, 81573254, 82173621].

Notes on contributors

Wei Ye

Wei Ye is a Postgraduate Student of Biostatistics at the Department of Health Statistics, College of Preventive Medicine, Army Medical University, Chongqing, China.

Ling Zhang

Ling Zhang is a Senior laboratory Technician at the Department of Health Education , College of Preventive Medicine, Army Medical University, Chongqing, China.

Wenqing Zhang

Wenqing Zhang was a Undergraduate Student of Biostatistics at the Department of Health Statistics, College of Preventive Medicine, Army Medical University, Chongqing, China.

Xiaojiao Wu

Xiaojiao Wu was a Postgraduate Student of Biostatistics at the Department of Health Statistics, College of Preventive Medicine, Army Medical University, Chongqing, China.

Dong Yi

Dong Yi is a Professor at the Department of Health Statistics, College of Preventive Medicine, Army Medical University, Chongqing, China.

Yazhou Wu

Yazhou Wu is a Director and Professor at the Department of Health Statistics, College of Preventive Medicine, Army Medical University, Chongqing, China.