To enable the above, we have developed approaches over the past few years with the following aims:
(A) to ensure optimal experimental design and appropriate data analysis based on knowledge of data structure. We have applied our research to many quantitative proteomics platforms and created supporting software and analysis pipelines.
(B) to enable the assignment of proteins to subcellular niches and multi protein complexes using quantitative proteomics (both targeted and non-targeted workflows) coupled with machine learning methods for data mining.
Robust experimental design is of paramount importance in any study involving quantitative proteomics. There is a bewildering number of experimental schema which could be applied to any study. We have studied optimal designs for many approaches including gel based methods (Karp and Lilley, Proteomics 2007), isobaric tagging (Karp et al, Mol Cell Proteomics 2010) and metabolic labeling (Russell and Lilley, Proteomics 2012). We are working with the Ralser group to determine optimal designs for label free proteomics experiments.
More recently we have developed iSPY, in collaboration with Sarah Martin and Thierry LeBihan at the University of Edinburgh which includes support for isobaric tagging and metabolic labeling (15N and SILAC) data. We have developed software to enable transparent analysis of data from multiple mass spectrometry platforms. In collaboration with Conrad Bessant (URL) we developed iTracker (Shadforth et al, BMC Genomics 2005) for use with isobaric tagged dataset and assisted in the creation of MRMaid, to assist in design of targeted proteomics data MRMaid, the web-based tool for designing multiple reaction monitoring (MRM) transitions (Mead et al, Mol Cell Proteomics 2009).
Robust statistical and computational data analysis is of vital importance to the above techniques, and to proteomics in general, to ensure that data sets are efficiently mined and do not contain unacceptable levels of false discovery. The Computational Proteomics Unit within CCP, develops bioinformatics and statistical tools, that utilize pattern recognition and machine learning methods to enable robust analysis of organelle proteomics and multi-protein complex data (Breckels et al, J.Proteomics 2013). The output of this research is manifested in the creation of open-source software solutions for quantitative data analysis that are applicable to the majority of quantitative proteomics applications.
Spatial Proteomics and Protein-Protein Interactions
Localisation of Organelle Proteins using Isotope Tagging (LOPIT) (Dunkley et al, Proc Natl Acad Sci USA 2006), which allows the assignment of proteins and protein complexes to sub-cellular locations, has been applied successfully to several biological systems (Nikolovski et al, Plant Physiol. 2012, Tan et al, Proteome Res. 2009, Hall et al. Mol Cell Proteomics. 2009). The ability to assign individual proteins accurately to specific sub-cellular structures and monitor their movement within cells is of paramount importance to our understanding of cellular mechanisms.
Interactomes using Parallel Affinity Capture (iPAC) is a method developed in collaboration with the St Johnston (Gurdon Institute) and Russell (Genetics Dept.) groups to determine genuine residents of multi protein complexes (Rees et al, Mol Cell Proteomics. 2011).
New Horizons for CCP
- The Lilley group has recently been awarded a sLoLa (BBSRC) in collaboration with the Russell and Martinez Arias groups (UCAM, Genetics), the Orengo and Jones groups (UCL) and Hubbard group (University of Manchester) to create a spatio-temporal map of the developmental fly interactomes.
- With the Martinez Arias group in the Department of Genetics, we are mapping spatio temporal changes in proteins associated with cell fate and differentiation.
There are currently no results for this search.