Everyone said that only a very small proportion of SNPs are functional. Is there anyone that can recommend some references on this topic, to discuss or analyze the proportion of functional SNPs against total SNPs?
SNPs, could be either synonymous (change in nucleotide but does not change the amino acid) or nonsynonymous (change in nucleotide and change in amino acid) . There are computational tools such as SIFT, Polyphen, PANTHER, i-mutant which can predict the effect of nonsynonymous SNPs as deleterious or not. You can refer the following links for further details.
About a year ago,I generated a brief summary of SNPs location from our SNPs database of more than 66 Million SNPs from 136 strains of laboratory mice, drawn from several sources - http://cgd.jax.org/cgdsnpdb/
a. 59% (38,919,578) of SNPs were found in Intergenic regions - gene free regions
b. 41% ( 27,109,231) of SNPs were located on genes (transcripts)
of the 41% of SNPs on genes, only 3.8% ( 1,102,004) were found on exons the rest on Introns. Now of the 1,102,004 SNPs on exons, 699,677 were on UTR and the rest were coding (synonymous: 290,014 or nonsynonymous: 161,260) - Given the location of functional elements along the genome, you can do the math