Using KING to infer kinship

[DonJoe]

Greetings!

I'm trying to use the software KING (http://people.virginia.edu/~wc9c/KING/) to check kinship between exome sequence samples. I'm not getting the expected results so I think I might be doing something wrong.

I have followed the procedure posted here: http://cphg.virginia.edu/quinlan/?p=300 to run the progran. My VCF file is created by running GATK UnifiedGenotyper with the options --dbsnp and --output_mode EMIT_VARIANTS_ONLY

The resulting output from KING show negative values for the kinship coefficient, although I know some of the individuals are siblings. Should I generate my VCF file in some other manner, like generate calls at all sites? This produce huge outputs though.

Has anyone else used KING or done similar analysis before? Any help or hints is greatly appreciated!

----------------------

These are the GATK options I've used:

-T UnifiedGenotyper \
-R ref.fa \
-I inputbams.list \
-o GATK_snpcall.vcf \
--output_mode EMIT_VARIANTS_ONLY \
--validation_strictness LENIENT \
-dcov 200 \
-l INFO \
-nt 8 \
--dbsnp dbsnp_135.b37.vcf

[Zigster]

Yeah, I tried Aaron's recipe once and got very low numbers (although siblings were higher than strangers). I think the no-calls is screwing it up.