Protein raftophilicity is the affinity of proteins for lipid ‘rafts’. Rafts denote nano- and submicro-sized biomembrane domains that are enriched in cholesterol and sphingolipids. These domains are considered relevant for maintaining specialized structures that constitute suitable sites for bioprocesses. Protein raftophilicity depends on features such as lipidation and GPI-anchoring. Can this affinity be inferred solely by knowing such features, without knowing the physical and physico-chemical properties of biomembranes? We tried to answer the question by an artificial neural network (ANN)-based bioinformatics approach. The ANN was trained to recognize feature-based patterns in proteins that are considered to be associated with lipid rafts. The trained ANN was then used to predict protein raftophilicity. We found that, in the case of α-helical membrane proteins, their hydrophobic length does not affect their raftophilicity. This is in agreement with confocal microscopy experiments on DOPC/SM/cholesterol bilayers with reconstituted model peptides, P-23 and P-29.
|Number of pages||1|
|Publication status||Published - 2015|
|Event||First Annual Danish Bioinformatics Conference - Odense, Denmark|
Duration: 27 Aug 2015 → 27 Nov 2015
|Conference||First Annual Danish Bioinformatics Conference|
|Period||27/08/2015 → 27/11/2015|