Comparison of Normalisation Methods for Mass Spectrometry Data
Author Information
Author(s): Wouter Meuleman, Judith YMN Engwegen, Marie-Christine W Gast, Jos H Beijnen, Marcel JT Reinders, Lodewyk FA Wessels
Primary Institution: The Netherlands Cancer Institute
Hypothesis
It is unclear which normalisation method yields the most informative result for SELDI-TOF mass spectrometry data.
Conclusion
Normalisation of mass spectra generally improves data quality, with methods based on data dispersion outperforming those based on central location.
Supporting Evidence
- Most normalisation methods significantly improved data quality compared to no normalisation.
- Methods using data dispersion for scaling outperformed those using central location.
- Local estimation of scaling factors provided additional performance improvements.
Takeaway
This study looked at different ways to clean up messy data from mass spectrometry, and found that some methods work much better than others.
Methodology
The study systematically compared 16 normalisation methods using multiple datasets and configurations to evaluate their performance in reducing variance and improving classification.
Potential Biases
Some methods may degrade signal quality while reducing variance, leading to biased results.
Limitations
The performance of normalisation methods may vary depending on the specific datasets and classifiers used.
Participant Demographics
The study involved serum samples from patients with colorectal cancer and healthy controls.
Statistical Information
P-Value
p = 2.05 × 10-67
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website