Abstract: Objective · To investigate the signalling pathways mediatedsynaptic cell adhesion molecules (sCAMs) in the process of human prenatal synapse formation. Methods · The single-cell RNA-sequencing dataset of prenatal human prefrontal cortex was downloaded GEO (Gene Expression Omnibus) database. The gene dynamics was modelled with pseudotime ordering approach and the protein-protein interaction (PPI) network was constructedutilizing gene co- analysis and PPI database. The interacting molecules and associated pathways of sCAMs were explored. Results · The gene dynamics of early synapse formation in excitatory neurons can be modelled with linear trajectory. PPI network analysis identified the interacting molecules of neurexins, neuroligins, and LAR-type receptor-type protein tyrosine phosphatases (LAR-type RPTPs). Guanine nucleotide exchange factor 9 (ARHGEF9) interacted with neurexins and neuroligins, while cell division cycle 42 (CDC42) was the hub of the network. Amyloid precursor protein (APP) interacted with neuroligins and leucine-rich repeat transmembrane neuronal protein 3 (LRRTM3), which is a ligand of neurexins. Finally, mitogen-activated protein kinase 8 (MAPK8), dual specificity phosphatase 4 (DUSP4), and CDC42, which participate MAPK signalling pathways, were involved in the PPI network of protein tyrosine phosphatase receptor type D (PTPRD, a member of LAR-type RPTPs) and its ligands leucine rich repeat and fibronectin type Ⅲ domain containing 1 (LRFN1), LRFN2, and LRFN5. Conclusion · Interacting proteins and associated pathways of neurexins, neuroligins, and LAR-type RPTPs can be predicted with bioinformatics methods, which may provide insights in experimental studies.

Key words: bioinformatics, synapse formation, neurodevelopment, RNA-sequencing, synaptic cell adhesion molecule (sCAM), protein-protein interaction