Biases in Transcriptomic Data Mining for Cancer Research
Author Information
Author(s): Hengrui Liu, Yiying Li, Miray Karsidag, Tiffany Tu, Panpan Wang
Primary Institution: Cancer Research Institute, Jinan University, Guangzhou, China
Hypothesis
This review aims to highlight the technical and biological biases in transcriptomic data mining for cancer research.
Conclusion
Understanding and mitigating biases in transcriptomic data analysis is crucial for accurate interpretation and reliable clinical outcomes in cancer research.
Supporting Evidence
- Transcriptomic data provides crucial insights into gene expression patterns and disease mechanisms.
- Technical biases include issues related to microarray, RNA-seq, and nanopore sequencing methods.
- Biological biases arise from factors like tumor heterogeneity and sample purity.
- Misinterpretations often occur when correlational data is erroneously assumed to imply causality.
Takeaway
This study talks about the problems researchers face when using data about genes in cancer studies, like mistakes that can happen when analyzing the data.
Methodology
The paper reviews existing literature on biases in transcriptomic data analysis and discusses various technical and biological factors affecting data interpretation.
Potential Biases
Potential biases include misinterpretation of correlation as causation and overlooking the complexity of tumor heterogeneity.
Limitations
The review does not provide new experimental data but rather synthesizes existing knowledge on biases in transcriptomic data.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website