Pancreatic ductal adenocarcinoma (PDAC) has an extremely poor prognosis. Understanding the specific mechanisms by which the tumour evades immune control, and how these mechanisms may be disrupted is critical to developing targeted immunotherapies. We developed novel computational approaches applied to single-cell multi-omic data (scRNA-seq, CITE-seq, TCR-/BCR-seq) from matched tumour-infiltrated CD45+ cells and peripheral blood in 12 patients, generating the most detailed single cell analysis of tumour-associated immune cells coupled with re-analysis of two publicly available datasets. This reveals a complex immune infiltrate showing patients have either a myeloid-enriched (ME) or adaptive enriched (AE) tumour microenvironment (TME). The level of lymphocyte enrichment is intrinsically linked with highly distinct B and T cell clonal selection, diversification, differentiation and function that impact pro- and anti-tumour mechanisms, and is associated with differential lymphocyte chemokine receptor upregulation upon tumour entry. Using TCR data, we see the largest clonal expansions in CD8 EM, CD8 senescent cells, and highly activated Treg populations that are induced within the TME from naïve cells and are supported through interactions with suppressive myeloid cells. ME tumours are dominated by cellular networks of myeloid-T cell interactions while AE tumours are dominated by T-B cell interactions. We identify chemokine pathways that potentially lead to a suppressive TME by both immune and non-immune cells, including targets of immunotherapy CCR8, CCR4 and CXCR4 and their ligands, as well as checkpoints TIGIT/PVR and SIRPA/CD47. Our novel approaches identify multiple novel therapeutic targets that should form the basis for rational design of a new generation of clinical trials in PDAC including adoptive cell therapy, boosting of B cell responses and specific Treg depletion approaches.