Adding ADAMContig back to ADAMVariant.

[fnothaft]

This PR modifies some code that was changed in #238/#243. Specifically, we add the ADAMContig object back to the ADAMVariant, and we make several of the fields in the ADAMVariant/ADAMGenotype objects nullable.

[AmplabJenkins]

All automated tests passed.
Refer to this link for build results: https://amplab.cs.berkeley.edu/jenkins/job/ADAM-prb/324/

[massie]

Thanks, Frank!

[hammer]

For future reference, why was ADAMContig removed from ADAMVariant, and why is it okay to add it back?

[fnothaft]

@hammer Good question; I'm not 100% sure why it was removed (paging @mlinderm) but I believe the reason it was removed is because the code that converts from VCF only has (easy) access to the contig name.* I've added it back, because it is useful to track this info if we are generating VCFs from data that is already in ADAM (and that already has all the contig data). Otherwise, the contig name is a subset of the full contig info, so there's no correctness issue.

* I believe that the full contig info can be recovered from the VCF input; due to the way that Picard is designed, you need to do some dirty dancing. I'll try to fix this in a follow-on pull request.