Effect of dbSNP build differences in variant calling

Question

[Mehulsharma.253]

I want to call variants on pre-aligned CRAM files for a 1000genomes sample, available for download. However, the dbSNP build used to recalibrate base quality scores (the dataset is already recalibrated) for all 1000genomes phase 3 pipelines is dbSNP build 142 which is the 2014 version (current build is 151)

Should I nevertheless proceed with calling variants despite the build differences? Would you recommend re-recalibrating the CRAM (after conversion to BAM) with the latest dbSNP build?