The Choice of the Filtering Method in Microarrays Affects the Inference Regarding Dosage Compensation of the Active X-Chromosome
2011

Impact of Filtering Methods on X-Chromosome Dosage Compensation Analysis

Sample size: 1467 publication Evidence: high

Author Information

Author(s): Raphaële Castagné, Maxime Rotival, Tanja Zeller, Philipp S. Wild, Vinh Truong, David-Alexandre Trégouët, Thomas Munzel, Andreas Ziegler, François Cambien, Stefan Blankenberg, Laurence Tiret

Primary Institution: INSERM UMRS 937, Pierre and Marie Curie University (UPMC, Paris 6) and Medical School, Paris, France

Hypothesis

Does the choice of filtering method in microarray studies affect the inference regarding dosage compensation of the active X-chromosome?

Conclusion

The method used for filtering lowly expressed genes in microarrays significantly impacts the conclusions about dosage compensation of X-linked genes.

Supporting Evidence

  • The study used data from a large-scale expression study in human monocytes.
  • Filtering methods significantly affected the estimated X∶AA ratio.
  • Using a uniform filtering threshold led to an over-estimation of X-linked expressions.
  • The hypothesis of dosage compensation was rejected when using chromosome-specific filtering.
  • Results were consistent across different tissues analyzed.

Takeaway

This study shows that how we filter gene data can change our understanding of how genes on the X chromosome are expressed compared to other genes.

Methodology

Microarray expression data from circulating monocytes in 1,467 individuals were analyzed using two different filtering methods to compare X-linked and autosomal gene expressions.

Potential Biases

Potential bias arises from using a uniform detection threshold that may disproportionately exclude lowly expressed genes on the X chromosome.

Limitations

The study's conclusions are dependent on the biological assumptions regarding the proportion of actively expressed genes on the X chromosome compared to autosomes.

Participant Demographics

1,467 unrelated subjects (51.1% men) of European descent aged 35–74 years.

Statistical Information

P-Value

p<10−31

Confidence Interval

95% bootstrap confidence intervals

Statistical Significance

p<10−31

Digital Object Identifier (DOI)

10.1371/journal.pone.0023956

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication