Intrahepatic cholangiocarcinoma (iCCA) is a rare and universally lethal malignancy arising in the liver. Incidence has steadily been increasing globally yet effective therapeutics are lacking. Currently, gold standard treatment is complete resection for which a minority of patients are suitable and recurrence rates post resection are high. For the majority of patients unsuitable for surgery, standard chemotherapeutic options extend life by only 3-6 months on average. As such there is a pressing need to elucidate the genetic and molecular features of this malignancy so as to inform the development of more effective therapies. To date, a small but robust set of driver genes have been discovered and for some targeted therapies in development, with mixed results in clinical trials. This work aimed to extend the set of known driver genes by exploring the long tail of infrequently mutated genes in patient sequencing data. Through computational prediction and in vivo screening, a set of driver genes were uncovered which cooperate with the more commonly encountered oncogenic RASG12 isoforms of KRAS and NRAS. Validation studies in vivo confirmed that the membrane cytoskeleton adaptor protein Merlin, encoded by NF2, and the canonical transmembrane semaphorin receptor PLXNB2 encoded by PLXNB2 are tumour suppressors occurring at low frequency in the patient population. Mutation of these genes significantly accelerated disease progression and propagated aggressive, invasive tumour phenotypes.