GTI: A Novel Algorithm for Identifying Outlier Gene Expression Profiles from Integrated Microarray Datasets
2011

New Algorithm for Identifying Outlier Gene Expression in Cancer

Sample size: 392 publication 10 minutes Evidence: high

Author Information

Author(s): Mpindi John Patrick, Sara Henri, Haapa-Paananen Saija, Kilpinen Sami, Pisto Tommi, Bucher Elmar, Ojala Kalle, Iljin Kristiina, Vainio Paula, Björkman Mari, Gupta Santosh, Kohonen Pekka, Nees Matthias, Kallioniemi Olli

Primary Institution: FIMM - Institute of Molecular Medicine Finland, University of Helsinki

Hypothesis

Can a new bioinformatic method effectively identify outlier genes in cancer datasets?

Conclusion

The Gene Tissue Index (GTI) algorithm successfully identifies potential oncogene outliers and drug targets in glioblastoma.

Supporting Evidence

  • The GTI identified 29 novel outlier genes in glioblastomas, including TYMS and CDKN2A.
  • 90% of the validated outlier genes showed a typical outlier staining pattern in glioblastoma samples.
  • The GTI performed equally well compared to existing methods in identifying outlier genes in simulated datasets.
  • The GTI reduced the number of false positives compared to other methods like COPA.
  • The GTI is particularly effective in heterogeneous datasets with varying sample sizes.

Takeaway

Researchers created a new tool to find special genes that are more active in cancer than in normal tissues, which could help in developing new treatments.

Methodology

The GTI algorithm was developed and validated using simulated data and real clinical datasets, comparing its performance with existing methods.

Potential Biases

Existing methods may identify false positives, such as genes that are highly expressed in both normal and cancer tissues.

Limitations

The GTI may not perform well when the proportion of outliers is very small.

Participant Demographics

The study analyzed gene expression data from 392 normal CNS tissue samples and 427 glioma samples.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1371/journal.pone.0017259

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication