Increased body mass index (BMI) increases risk of chronic diseases including heart disease, stroke, and type 2 diabetes (T2D). Through genome-wide association studies (GWAS), 97 genetic variants, mostly common (MAF>5%) and non-coding, have been identified for BMI. To investigate the role of coding variants with putative functional effects (missense, nonsense, splicing, stop gain), we meta-analyzed ExomeChip data of 526,508 individuals from over 110 studies, predominantly of European ancestry. Individual studies performed association analyses using Rvtest or RareMetalWorker; results were subsequently combined in single variant meta-analyses and gene-based tests (SKAT, VT) stratified by ancestry using RareMETALS. Gene-based tests were grouped based on functional annotation and in silico predicted impact on proteins for variants with a MAF<5%. Among the 245,497 variants analyzed, of the 93 loci that reached array-wide significance (P<2E-07); 56 were novel or independent of previously identified BMI loci. This includes low frequency variants in ACHE (rs1799805, p.H353N, p=8.01E-10, EAF=4%, β(SE)=0.032(0.005) SD/allele), HIP1R (rs34149579, p.V67F, p=1.88E-08, EAF=4%, β(SE)=-0.027(0.005) SD/allele) and MC4R (rs13447324, p.Y35X, p=2.37e-11, EAF=0.01%, β(SE)=0.64(0.09) SD/allele). Eight gene-based associations were significant (p<2.5E-06), including GIPR (p=7.17E-09), whose effects were independent of the common GWAS proxy in the gene (rs1800437, p. E354Q, p=7.91E-30, MAF=20%, β(SE)=-0.029(0.003) SD/allele). GIPR is part of the GLP1 pathway responsible for the secretion of postprandial insulin. Other gene-based associations were largely driven by a single variant, such as for KSR2 (p=7.15E-09), RAPGEF3 (p=8.91E-15) and PRKAG1 (p=2.75E-12). Ksr2 knockout mice are obese. Multiple variants in KSR2 have been shown to associate with severe early onset obesity. Mutations in KSR2 disrupt the Raf-MEK-ERK pathway, resulting in impaired fatty acid oxidation and glucose oxidation. RAPGEF3 is involved in the GLP1 pathway and regulates glucose sensitivity of the KATPchannel, which regulates insulin secretion. PRKAG1 is involved in AMPK signaling and has previously been associated with T2D and weight gain on antipsychotic medication. Taken together, by interrogating coding variants in a large sample set, we have identified both additional genes and variants associated with BMI, some of which may be causal.