Pattern and cluster analysis

The purpose of most pattern detection methods is to represent the variation in a data set into a more manageable form by recognizing classes or groups. The data typically consist of a set of objects described by a number of characters. In microarray data, for example, an object is a gene, while a character is the expression of that gene under a particular condition.

If the objects were always described by only two or three characters, there would not be much need for pattern detection methods. Just plotting the data in two or three dimensions, respectively, would be sufficient to distinguish groups. However, typically, objects are characterized by more than three characters, so simply plotting the data is not possible. Other ways need to be found to represent the data. There are basically two approaches that can be taken:

• Principal component analysis, which reduces the number of characters by finding two or three new characters that are combinations of the old characters
  
• Cluster analysis, which reduces the number of objects by placing them into groups

The book chapter Theoretical Aspects of Pattern Analysis gives a simple introduction to both principal component and cluster analysis. Applications of both types of analyses can be found in the papers listed below.

• Novel candidate genes associated with hippocampal oscillations
  Jansen, R., Timmerman, J., Loos, M., Spijker, S., Van Ooyen, A., Brussaard, A. B., Mansvelder, H. D., The Neuro-Bsik Mouse Phenomics Consortium, Smit, A. B., De Gunst, M., and Linkenkaer-Hansen, K. (2011). PLoS ONE 6(10): e26586. doi:10.1371/journal.pone.0026586.
  [Abstract] [Full text: PDF]

• Inbred mouse strains differ in multiple hippocampal activity traits
  Jansen, R., Linkenkaer-Hansen, K., Heistek, T., Timmerman, J., Mansvelder, H. D., Brussaard, A. B., De Gunst, M., and Van Ooyen, A. (2009). Eur. J. Neuroscience 30: 1092-1100.
  [Abstract] [Full text: PDF]

• Theoretical aspects of pattern analysis
  Van Ooyen, A. (2001). In: Dijkshoorn, L., Towner, K. J., and Struelens, M., eds. New Approaches for the Generation and Analysis of Microbial Typing Data. Amsterdam: Elesevier, pp. 31-45.
  [Abstract] [Full text: PDF]

• Endemic acinetobacter in intensive care units: epidemiology and clinical impact
  Dijkshoorn, L., Van Dalen, R., Van Ooyen, A., Bijl, D., Tjernberg, I., Michel, M. F., and Horrevorts, A. M. (1993). J. Clin. Pathol. 46: 533-536.
  [Abstract] [Full text: PDF]

• The allopolyploid origin of Sedum rupestre subsp. rupestre (Crassulaceae)
  't Hart, H., Sandbrink, J. M., Csikos, I., Van Ooyen, A., and Van Brederode, J. (1993). Pl. Syst. Evol. 184: 195-206.
  [Abstract] [Full text: PDF]

• Cellular composition of germinal centers in lymph nodes after HIV-1 infection: evidence for an inadequate support of germinal center B lymphocytes by follicular dendritic cells
  Rademakers, L. H. P. M., Schuurman, H.-J., De Frankrijker, J. F., and Van Ooyen, A. (1992). Clinical Immunology and Immunopathology 62: 148-159.
  [Abstract] [Full text: PDF]

• Ultrastructural analysis of germinal centres in lymph nodes after HIV-1 infection
  Rademakers, L. H. P. M., Schuurman, H.-J., De Frankrijker, J. F., and Van Ooyen, A. (1990). Micron and Microscopica Acta 21: 267-268.
  [Full text: PDF]

• Comparison of clinical acinetobacter strains using a carbon source growth assay
  Dijkshoorn, L., Van Ooyen, A., Hop, W. C. J., Theuns, M., and Michel, M. F. (1990). Epidemiol. Infect. 104: 443-453.
  [Abstract] [Full text: PDF]

• Iterative character weighting based on mutation frequency: a new method for constructing phyletic trees
  Van Ooyen, A., and Hogeweg, P. (1990). J. Mol. Evol. 31: 330-342.
  [Abstract] [Full text: PDF]