I have a VCF file from a set of exomes (about 50) and would like end up with a set of likely haplotypes for a small set of regions (genes). I imagine using the LD information in projects like hapmap/1000genomes, this should be approximately possible. We could also use read-based haplotype methods, but with exomes, I don't expect this to be a fruitful exercise.
We are ultimately interested in how germline variation might impact drug sensitivity in our samples, so we want to reduce the number of potential variants to a minimum consistent set.
I have phased a small population (40 individuals) using Beagle. Do you want to impute missing genotypes? WIll you be comparing this with 1kg as a background? The amount of LD will be tricky. Brian Browning's review may help you? http://www.nature.com/nrg/journal/v12/n10/pdf/nrg3054.pdf?WT.ec_id=NRG-201110