In vcf2bed, we convert from 1-based, closed [start, end] Variant Call Format v4 (VCF) to sorted, 0-based, half-open [start-1, end) extended BED data (cite).

For your example, a single-base variant at position 89017961 would map to 89017960-89017961, by default.

(Other custom variant process options are available to handle the coordinates in a different fashion; see the documentation for information about the --snvs, --insertions and --deletions command-line options.)

If you plan to integrate your data with other UCSC-formatted BED datasets, consistently using the prefix chr for chromosome names is a good idea, especially if you plan to integrate toolkits like BEDOPS, GROK or Bedtools to process BED datasets, but there are other approaches you can take, depending on the data or the lab or institution you're working with.

You can fix a lot of this stuff with standard UNIX piping. Building processing pipelines with UNIX pipes is a powerful option.

For example, convert to BED and look at the first few lines with head:

$ vcf2bed < foo.vcf | head
...

Then use awk or other tools of choice to modify fields with prefixes, remove capitalization, etc.

To demonstrate, you can prefix chromosome numbers with chr very easily:

$ vcf2bed < foo.vcf | awk '{ print "chr"$1"\t"$2"\t"$3; }' - > foo.fixed.bed

There is a tool kit for genome analysis especially for BED formats called BEDOPS. You can convert vcf to bed using these convertor tools available in BEDOPS.