NEW YORK (GenomeWeb News) – Synonymous SNPs that don't change the amino acid sequence encoded by a gene appear just as likely to influence human disease as non-synonymous SNPs that do, according to a paper appearing online recently in PLoS ONE by researchers from Stanford University and the Lucile Packard Children's Hospital.
"Most people ignore synonymous SNPs because they think that what could the importance be if it's not changing the amino acid in a protein," senior author Atul Butte, a biomedical informatics and pediatrics researcher at Stanford, told GenomeWeb Daily News.
"But at the same time, there are a lot of SNPs that show up in GWAS," he added, noting that such SNPs have been detected in coding regions as well as gene deserts.
Butte and his team brought together findings from some 2,100 studies to come up with a database representing more than 21,000 documented associations between SNPs and disease. When they scrutinized the frequency and effect sizes for nine different types of SNPs in this database, the researchers discovered that about as many synonymous genetic variants have been tied to disease as non-synonymous SNPs — with comparable effect sizes.
Although recent genome-wide association studies have uncovered thousands of SNPs associated with various common diseases, the researchers explained, it's unclear whether non-coding SNPs and synonymous coding SNPs have the same likelihood of influencing disease as non-synonymous coding SNPs that drastically alter a protein's amino acid sequence.
To explore this question, Butte and his co-workers incorporated and annotated data on 21,429 SNP-disease associations reported in 2,113 published studies. The data is housed in a database that also contains details about the disease being assessed, phenotype, population, and more.
Using this information, the researchers were able to compare the frequency of different sorts of SNPs and their odds ratios, looking specifically at nonsense, non-synonymous, and synonymous SNPs, as well SNPs found in various parts of the genome, including 5'-UTRs, 3'-UTRs, UTR-adjacent areas and intronic and intergenic regions.
While nonsense SNPs producing premature stop codons were most often linked to disease, the team explained, they found similar association frequencies for non-synonymous and synonymous SNPs and for SNPs in 5'-UTR regions of the genome. They also found similar effect sizes for the synonymous and non-synonymous variants.
The team's analyses also indicated that variants in microRNA binding site sequences and other 3' untranslated regions may be an unappreciated source of disease-associated variants. In addition, they noted that SNPs in intronic parts of the genome may be relevant to disease — particularly those in the first intron, which appear to be more likely to harbor disease-associated variants than other introns.
Though they conceded that the synonymous SNPs don't necessarily represent causal variants, the researchers predict that focusing on nonsense and non-synonymous SNPs alone when doing functional studies following up on GWAS would overlook more than 90 percent of the variants linked to disease.
"Our results suggest that [synonymous SNPs] are just as likely to be involved in disease mechanisms, so we recommend that [synonymous SNPs] discovered from GWAS should also be examined with functional studies," the team wrote.
More research is needed to determine how synonymous SNPs could contribute to disease, Butte explained. But past studies suggest variants altering the type of codon designated by sequence can lead to changes in transfer RNAs binding, he noted, and synonymous changes could also affect RNA structure and/or stability.
"There are these theories, but it's really case examples right now," Butte said. He noted that his team is currently doing follow-up work looking at the role of specific synonymous SNPs in disease.