Developing an object-based colocalisation analysis method to measure synaptic diversity
Item statusRestricted Access
Embargo end date06/07/2020
MetadataShow full item record
Protein colocalisation is of particular importance in the study of protein function. To address the inadequacies of previous colocalisation analysis methods, the novel Vicinity-based Localisation Adjacency Determination (VLAD) object-based colocalisation analysis method was developed. VLAD provides three main colocalisation measurements: the proportion of colocalising objects in a dataset, the probability of true colocalisation for individual objects, and the spatial relationship (distance) between colocalising objects. VLAD, validated by extensive testing in simulated data in a wide range of conditions (localisation densities, levels of colocalisation and colocalisation distances), was shown to outperform the state-of-the-art colocalisation analysis method SODA (Statistical Object Distance Analysis). VLAD was used to study the distribution and colocalisation of three key synaptic proteins: GluN1 (obligatory subunit of NMDA receptors), PSD95 and SAP102 (scaffolding proteins at excitatory synapses). In total, over 62.5 million puncta or puncta assemblies of these proteins were analysed in the mouse hippocampus during early development, making this the largest triple colocalisation brain mapping study of this sort. GluN1, PSD95 and SAP102 associate in a combinatorial fashion, giving rise to 7 synaptic protein punctum subtypes. The subtype compositions of the hippocampal subregions diverge in development and the differences in subtype compositions in the adult hippocampus may underlie the distinct functions performed by each component of the hippocampal circuit. It was found that a high proportion of the puncta of each protein were non-colocalising in the adult mouse – 67% of GluN1, 48% of PSD95 and 27% of SAP102. Interestingly, NMDA receptors (GluN1) appear to colocalise with PSD95 only in the presence of SAP102, hinting at a possible codependence between these proteins. This study demonstrated the potential of VLAD in the field of brain mapping.