Ensemble-based data assimilation and the localisation problem

The “butterfly effect” is a popularly known paradigm; commonly it is said that when a butterfly flaps its wings in Brazil, it may cause a tornado in Texas. This essentially describes how weather forecasts can be extremely senstive to small changes in the given atmospheric data, or initial conditions, used in computer model simulations. In 1961 Edward Lorenz found, when running a weather model, that small changes in the initial conditions given to the model can, over time, lead to entriely different forecasts (Lorenz, 1963). This discovery highlights one of the major challenges in modern weather forecasting; that is to provide the computer model with the most accurately specified initial conditions possible. A process known as data assimilation seeks to minimize the errors in the given initial conditions and was, in 1911, described by Bjerkness as “the ultimate problem in meteorology” (Bjerkness, 1911).

Viral nucleolar localisation signals determine dynamic trafficking within the nucleolus

Localisation of both viral and cellular proteins to the nucleolus is determined by a variety of factors including nucleolar localisation signals (NoLSs), but how these signals operate is not clearly understood. The nucleolar trafficking of wild type viral proteins and chimeric proteins, which contain altered NoLSs, were compared to investigate the role of NoLSs in dynamic nucleolar trafficking. Three viral proteins from diverse viruses were selected which localised to the nucleolus; the coronavirus infectious bronchitis virus nucleocapsid (N) protein, the herpesvirus saimiri ORF57 protein and the HIV-1 Rev protein. The chimeric proteins were N protein and ORF57 protein which had their own NoLS replaced with those from ORF57 and Rev proteins, respectively. By analysing the sub-cellular localisation and trafficking of these viral proteins and their chimeras within and between nucleoli using confocal microscopy and photo-bleaching we show that NoLSs are responsible for different nucleolar localisations and trafficking rates.

Localisation studies of cell-wall degrading enzymes in soybean

Postembedding immunoelectron microscopy has been used to investigate the diffusibility of an endo-beta-1,4-glucanase and a xylanase from A. niger in soybean. The results showed more specific localisation of the enzymes into the protein and lipid bodies of soybean cells. This was against our hypothesis that suggested that the enzymes should be localised in the cell wall.

Cell motility and SCAR localisation in axenically growing Dictyostelium cells.

Dictyostelium is a popular experimental organism, in particular for studies of actin dynamics, cell motility and chemotaxis. We find that the motility of axenic cells is unexpectedly different from other strains during growth. In particular, vegetative AX3 cells do not show detectable localisation of SCAR and its regulatory complex to actin-rich protrusions such as filopodia and pseudopodia. Similarly, a range of different mutations, in particular knockouts of members of the SCAR complex and Ras proteins, cause different phenotypes during vegetative growth in different parental strains. Development reverses this unusual behaviour; aggregation-competent AX3 cells localise SCAR in the same way as cells of other strains and species. Studies on cell motility using vegetative cells should therefore be interpreted with caution.