PI: Johan Trygg

Staff: Hans Stenlund, PhD student

Background

The overall goal of FuncFiber is to establish the basis for wood fiber improvement by unravelling the function of genes underlying the wood forming process in Poplar trees. It is well known that environmental and genetic changes in a biological sample are reflected in multiple developmental stages. Within the FuncFiber project, more than 2000 transgenic poplar trees will be characterised and analysed using a number of complementary bioanalytical platforms that reflect different developmental stages. These techniques can be divided into three main classes;

1. High-throughput profiling (e.g. FTIR, NIR, PyMS spectroscopy)
2. Semi-throughput profiling (e.g. metabolomics and microarray technology)
3. Low-throughput profiling (e.g. enzymology, FTIR-imaging and proteomics)

These techniques typically generate 1000-100000 data points (variables) for each biological sample. From this standpoint, it is clear that traditional statistical methods, that assume independent variables (i.e. more samples than variables), should not be used. In biology, chemometric methodology has been largely overlooked in favor of traditional statistics. It is not until recently that the overwhelming size and complexity of the "omics" technologies has driven biology towards the adoption of chemometric methods. This project will integrate and model information from multiple levels in a holistic attempt to understand the whole system. To be successful, chemometrics technology plays an integral part. The design of experiments methodology is crucial, and this requires the involvement and dedication of all groups within the centre.

There are still many challenging opportunities. One is concerned with sample collection and experimental protocols to enable comparisons between different profiling technologies. Another major challenge is how to utilize the information from multiple and different types of analytical platforms (e.g. NMR, PyMS and FTIR-imaging), that each provide unique fingerprints of the biological samples.

Aims

This project will develop tools and strategies on how to combine, model and predict biological information based on transgenic trees generated from multiple profiling platforms (e.g. NIR, FTIR-imaging, PyMS, metabolomics, enzymology and transcriptomics). This work includes;

• Using state of the art chemometric methods, e.g. the OPLS methodology, and where insufficient, new theories and methods will be developed.
• Design of Experiments (DoE) as an organised strategy in planning experiments and making samples and data informative. This encompasses optimizing experimental protocols, sample dynamics and normalisation procedures towards a combined profiling approach.
• Develop methods and strategies for analysis of NIR, VIS and FTIR-imaging.
• A strategy to construct "biologically equivalent samples" to be able to match the multivariate profile measured from each technology but having different developmental stages.
• Construction of a database, the FuncFiber 100 that contains all relevant information, data and results from the different analytical methods. The goal is to predict interesting fiber properties early on in the developmental process and to establish the link between genes and their function.