What tools do you recommend to predict the likely functional impact of silent SNVs (i.e. synonymous, UTR, intronic, and intergenic variants) in humans? I understand that this is much more difficult than for non-synonymous variants, but a prioritization based on sequence conservation (e.g. across closely related vertebrates) is probably sufficient as first-line approximation.
To my current knowledge, popular variant annotation tools (e.g. ANNOVAR, SnpEff/SnpSift, dbNSFP) provide (conservation) scores only for non-synonymous variants. SilVA (http://compbio.cs.toronto.edu/silva/) analyzes silent variants, but I shy away from trained classifiers because of unknown biases introduced by the training data set. I guess annotating input variants with sortable numeric conservation scores would suffice for my purpose.
EDIT: Found a recent review on this subject (Hunt RC et al., 2014; Exposing synonymous mutations), but did not spot any tool in there with a specific focus on sequence conservation.