A new holistic exploratory approach to Systems Biology by Near Infrared Spectroscopy evaluated by chemometrics and data inspection

Abstract

There is a need for an improved biological and theoretical interpretation of Near Infra‐Red Spectral (NIRS) fingerprints from tissues that could contribute with holistic overview to fine‐grained detail modelled in Systems Biology. The concept of gene expression in self‐organised networks was experimentally tested in a barley endosperm model with molecularly defined and undefined mutants. Surprisingly reproducible gene‐specific NIRS fingerprints were observed directly in log1/R MSC pre‐treated spectra that could not be accurately represented by destructive mathematical models. A mutant spectrum in an isogenic background represents the physiochemical expression of the gene in the whole network (tissue). The necessary holistic overview that is needed experimentally to introduce Ilya Prigogine's theory on self‐organisation in Systems Biology was supplied by defining the spectral phenome. Interval spectral information on genotypes and environment was classified by interval Extended Canonical Variates Analysis (iECVA). Genetic changes in spectra were interpreted by interval Partial Least Squares Regression (iPLSR) correlations to chemical variables. A new pathway regulation was detected. The finely grained ‘bottom up’ modelling of molecular and chemical data from pathways requires a coarsely grained exploratory ‘top down’ overview by NIRS to account for the outcome of self‐organisation. The amplification of expression from a gene to the phenome (pleiotropy) can now for the first time be quantified as a whole reproducible phenomenological pattern by NIRS and compared to other gene spectra. It explains published findings that transformed respectively mutated genes in genetically modified organisms (GMOs) and cancer patients can be detected unsupervised from tissues by spectroscopy, chemometrics and data inspection. Copyright © 2007 John Wiley & Sons, Ltd.

A physiochemical theory on the applicability of soft mathematical models—experimentally interpreted

“…The differential spectra of the two carbohydrate C mutants 16 and lys5.f in yellow-red ( Figure 6A) follow each other closely. The negative peak at 1930 nm in the water region indicates a mean difference in dry matter between C and NþP genotypes of 1.5% with a PLSR correlation coefficient of r ¼ 0.96 [24]. The spectra from the allelic protein P mutant's lys3.a, lys3.b, lys3c follow a typical P pattern different from that of the C spectra.…”

“…In Figure 6A (Material 1) the 'spectral s-deterministic fine-tuning' (Table I:2) of five differential mutant (field) spectra are shown normalised to the isogenic parent line Bomi that is shown as a straight line. They represent the spectral phenome of these mutants [15,24]. The differential spectra summarise the physiochemical effect of the cause of each mutant lesion on the expression of all other genes during seed synthesis.…”

“…The unique barley seed/endosperm mutant model represents the final physiochemical outcome (effect) of a single mutation cause in DNA in the self-organised endosperm tissue (definition in [15,24]. The main component of the barley seed, surrounded by husk and seed coat layers (approx 18%), is the starch rich endosperm tissue (80%) that is used as a food storage organ for future growth of the germ (2%).…”

“…The approximate increase from 5 to 20% in ß-glucan and the 50% decrease in starch [15] in the lys5.f mutant (Table II) is accompanied by a drastic increase in water content ( Figure 4E [17,26]) measured during seed synthesis from 16 to 39DAF with over 10 relative percentages in the lys5.f endosperm compared to the normal control. This condition is likely to change the water activity in stochastic (Table I:4) diffusive reactions throughout the endosperm tissue, which in turn will influence all chemical equilibriums, and enzyme activities in an unpredictable indeterministic way (Table I:6) [24]. The high osmotic pressure in plant cells is further likely to amplify the effect of a change in water activity.…”

Utilising NIRS for Qualitative and Non-destructive Identification of Seed Mutants in Large Populations

“…This involves the classification of samples according to their spectroscopic properties. Samples are clustered according to their similarity/dissimilarity (Munck 2007). This is based on "spectroscopic fingerprints" of each sample.…”