BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Centering\, scaling\, and transformations: improving the biologica
 l information content of metabolomics data - Kian Kai Cheng (Biochemsitry 
 Department\, Uni of Cambridge)
DTSTART:20070831T130000Z
DTEND:20070831T140000Z
UID:TALK7814@talks.cam.ac.uk
CONTACT:Dr N Karp
DESCRIPTION:A presentation and discussion of a paper by Robert A van den B
 erg et al titled 'Centering\, scaling\, and transformations: improving the
  biological information content of metabolomics data'\n\nDetails:BMC Genom
 ics 2006\, 7:142 doi:10.1186/1471-2164-7-142\n\nAbstract\n\nBackground:\nE
 xtracting relevant biological information from large data sets is a major 
 challenge in functional genomics research. Different aspects of the data h
 amper their biological interpretation. For instance\, 5000-fold difference
 s in concentration for different metabolites are present in a metabolomics
  data set\, while these differences are not proportional to the biological
  relevance of these metabolites. However\, data analysis methods are not a
 ble to make this distinction. Data pretreatment methods can correct for as
 pects that hinder the biological interpretation of metabolomics data sets 
 by emphasizing the biological information in the data set and thus improvi
 ng their biological interpretability.\n\nResults:\nDifferent data pretreat
 ment methods\, i.e. centering\, autoscaling\, pareto scaling\, range scali
 ng\, vast scaling\, log transformation\, and power transformation\, were t
 ested on a real-life metabolomics data set. They were found to greatly aff
 ect the outcome of the data analysis and thus the rank of the\, from a bio
 logical point of view\, most important metabolites. Furthermore\, the stab
 ility of the rank\, the influence of technical errors on data analysis\, a
 nd the preference of data analysis methods for selecting highly abundant m
 etabolites were affected by the data pretreatment method used prior to dat
 a analysis.\n\nConclusion:\nDifferent pretreatment methods emphasize diffe
 rent aspects of the data and each pretreatment method has its own merits a
 nd drawbacks. The choice for a pretreatment method depends on the biologic
 al question to be answered\, the properties of the data set and the data a
 nalysis method selected. For the explorative analysis of the validation da
 ta set used in this study\, autoscaling and range scaling performed better
  than the other pretreatment methods. That is\, range scaling and autoscal
 ing were able to remove the dependence of the rank of the metabolites on t
 he average concentration and the magnitude of the fold changes and showed 
 biologically sensible results after PCA (principal component analysis).\nI
 n conclusion\, selecting a proper data pretreatment method is an essential
  step in the analysis of metabolomics data and greatly affects the metabol
 ites that are identified to be the most important.\n
LOCATION:Meeting room 1 cambridge system biology centre
END:VEVENT
END:VCALENDAR