Error structure of spectroscopic data (NIR, FTIR etc) - and how to deal with them …. Harald Martens and Achim Kohler

Centre for Biospectroscopy and Data Modelling, Nofima Food, Ås, Norway CIGENE – Center for Integrative Genetics, University of Life Sciences, Ås, Department of Mathematical Sciences and Technology (IMT), Norwegian University of Life Sciences, Ås, Norway

1

My own field: Measurements and modelling in systems biology Environment, human activity DNA

mRNA

Sequencing, SNP, AFLP,

Realtime PCR Micro-array

Proteome

Metabolome

1D-, 2D Electrophoresis MALDI-TOF LC-MS

GC,LC (-MS)

Biological Structure NIR, FT-IR Raman Flourescence Serotyping

Other phenotypes Disease incidence Virulence Drug sensitivity Biofilm formation Sensory Science Economy

Data analysis: Integrating different types of bio-data Look for common variation patterns Make quantitative prediction and forecasting Identify outliers

2

Now the real fun starts: feed-back !

Environment, human activity DNA Sequencing, SNP, AFLP,

mRNA Realtime PCR Micro-array

Proteome

Metabolome

1D-, 2D Electrophoresis MALDI-TOF LC-MS

GC,LC (-MS)

Biological Structure NIR, FT-IR Raman Flourescence Serotyping

Other phenotypes Disease incidence Virulence Drug sensitivity Biofilm formation Sensory Science Economy

High-dimensional dynamic, non-linear ODEs Spatial PDEs Possible, since we how are getting relevant and reliable high-throughput, high-dimensional instrumentation

3

Biospectroscopy •

Wavelength ranges: – UV-Vis (2500 nm – Raman Scattering -”– Fluorescence: (mainly