To quote from a paper 'exome sequencing identifies the cause of a mendelian disorder'
"the clear majority of alleleic variants known to underlie Mendelian disorders disrupt protein-coding sequences. Splice acceptor and donor sites represent an additional class of sequences that are enriched for highly functional variation...a large fraction of rare non synonymous variants in the human genome are predicted to be deleterious. This contrasts with non coding sequences where variants are more likely to have neutral or weak effects on phenotype"
However when considering a complex multigenic disease such as diabetes I don't think it wold be prudent to focus on exonic variants.
This is also interesting about synonymous SNPs