14Esophageal Adenocarcinoma (EAC) is a poor prognosis cancer type with rapidly rising incidence. Our 15 understanding of genetic events which drive EAC development is limited and there are few molecular 16 biomarkers for prognostication or therapeutics. We have accumulated a cohort of 551 genomically 17 characterised EACs (73% WGS and 27% WES) with clinical annotation and matched RNA-seq. Using a 18 variety of driver gene detection methods we discover 65 EAC drivers (66% novel) and describe 19 mutation and CNV types with specific functional impact. We identify a mean of 3.7 driver events per 20 case derived almost equally from copy number events and mutations. We compare driver mutation 21 rates to the exome-wide mutational excess calculated using Non-synonymous vs Synonymous 22 mutation rates (dNdS). We see mutual exclusivity or co-occurrence of events within and between a 23 number of EAC pathways (GATA factors, Core Cell cycle genes, TP53 regulators and the SWI/SNF 24 complex) suggestive of important functional relationships. These driver variants correlate with tumour 25 differentiation, sex and prognosis. Poor prognostic indicators (SMAD4, GATA4) are verified in 26 independent cohorts with significant predictive value. Over 50% of EACs contain sensitising events for 27 CDK4/6 inhibitors which are highly correlated with clinically relevant sensitivity in a panel EAC cell 28 lines. 29 30 simplest of these features is the tendency of a mutation to co-occur with other mutations in the 48 same gene at a high frequency, as detected by MutsigCV 9 . MutsigCV has been applied on several 49 occasions to EAC cohorts 6,10,11 and has identified ten known cancer genes as high confidence EAC 50 drivers (TP53, CDKN2A, SMAD4, ARID1A, ERBB2, KRAS, PIK3CA, SMARCA4, CTNNB1 and FBXW7). 51 However these analyses leave most EAC cases with only one known driver mutation, usually TP53, 52 due to the low frequency at which other drivers occur. Equivalent analyses in other cancer types 53 have identified three or four drivers per case 12,13 . Similarly, detection of copy number driver events 54 in EAC has relied on identifying regions of the genome recurrently deleted or amplified, as detected 55 by GISTIC 10,14-17 . However, GISTIC identifies relatively large regions of the genome, containing 56 hundreds of genes, with little indication of which specific gene-copy number aberrations (CNAs) may 57 actually confer a selective advantage. There are also several non-selection based mechanisms which 58 can cause recurrent CNAs, such as fragile sites where a low density of DNA replication origins causes 59 frequent structural events at a particular loci. These have not been differentiated properly from 60 selection based recurrent CNAs 18 . 61Without proper annotation of the genomic variants which drive the biology of EAC tumours 62we are left with a very large number of events, most of which are likely to be inconsequential, 63 making it extremely difficult to detect statistical associations between genomic variants and various 64 biologi...