Inbreeding coefficients, Ho, He in PLINK software

[mira1]

I have calculated individual inbreeding coefficients in PLINK software ( --het) and took the average of all individuals from a population to calculate inbreeding in the different populations. These coefficients are based on loss of heterozygosity (i.e. based on observed versus expected number of homozygous genotypes).

Now, I have calulated observed and expected heterozygosity (Ho and He), also in PLINK software (--hardy), and for each population Ho is larger than He while in the table of individual inbreeding coefficients the number of observed homozygous genotypes is larger than the number of expected genotypes. So, this would mean that there are more heterozygous individuals than expected. How come? or do I interpreted the results in the wrong way? I also cannot find how these parameters are exactly measured and what's the difference between them.

I hope someone has an answer.

In addition, for both calculations I used a pruned dataset (--mind 0.1 --maf 0.05 --geno 0.1 --hwe 0.001 --indep-pairwise 100 25 0.2)

Edited January 6, 2014 by Miramariska