The generation of the face and neck is unique to vertebrate embryonic development, and requires a complex orchestration of dynamic molecular and tissue interactions. Elucidation of the molecular control of craniofacial development is important for the understanding of morphogenesis, evolution and disease. Branchial arches (BAs) are transient structures of early craniofacial development giving rise to different components of the oropharyngeal apparatus. BA development requires the contribution of endoderm, mesoderm, ectoderm and cranial neural crest cells (CNCCs). CNCCs form transiently during embryogenesis and are characterised by extensive multipotency and migratory abilities. CNCCs migration represents a key step in the evolution of the vertebrate head. This research aimed to uncover the molecular mechanisms regulating the mouse second branchial arch (IIBA), which contributes to the formation of the middle ear and neck. IIBA development depends on the expression of the transcription factor Hoxa2. Abnormal development of the IIBA is responsible for many craniofacial congenital malformations, therefore, dissecting the IIBA molecular control is important for gaining insights into development and disease mechanisms. Inference of regulatory networks from omic data has provided a valuable tool to study biological mechanisms. A network model from mouse IIBA transcriptomic data was generated by using an overlapping module algorithm. This model identified CBL, CTNNB1, EP300, EGFR and CDH1 (E-CADHERIN) as putative regulatory proteins associated with IIBA development and suggested the presence of an epithelial to mesenchymal transition (EMT) event. EMT is a crucial process occurring during embryonic development and adult tissue homeostasis, which is characterised by loss of cell-cell adhesion and increased cellular motility. A prerequisite for EMT is the loss of E-CADHERIN. By means of a network comparison approach, the study uncovered a statistically significant overlap between the IIBA network model and one generated from transcriptomic data of wild type (wtD3) versus E-cadherin knock-out (Ecad-/-) mouse embryonic stem cells (mESCs) (Ecad-/- vs wtD3), with candidate proteins shared between the two models. The function of the common putative regulator EP300 and its related histone epigenetic signature H3K27ac was interrogated in both wtD3 and Ecad-/- mESCs by chromatin immunoprecipitation and massive parallel sequencing (ChIP-seq) analysis. The results showed that EP300 and H3K27ac binding signatures were enriched in Ecad-/- mESCs, suggestive of a poised epigenetic EMT phenotype. EP300 and H3K27ac were increased at network candidate genomic loci. Conversely, ChIP-seq analysis on human induced pluripotent stem cells (hiPSCs) treated with an E-CADHERIN inhibiting peptide revealed absence of overlap with the network, suggesting they could map to a different stage of IIBA development. The function of the candidate molecules identified through the IIBA network model was tested by chemical inhibition assays in mESCs induced to differentiate with retinoic acid (RA) treatment and subsequent analysis of Hoxa2 transcript expression. CDH1 and CTNNB1 repress Hoxa2 transcript expression, whereas CBL and EGFR are shown to be required for Hoxa2 activation. This research is the the first demonstration of a network biology approach to the study of IIBA development in silico and confirmation of candidate protein function using an in vitro model of IIBA development.