The ‘missing heritability problem’ is the inability to tag all the genetic variance of a trait using genome wide single nucleotide polymorphism. Here, we compute missing heritability for two populations of cattle phenotyped for ten tropical adaptation traits that exhibit variable genetic architectures. We derived genomic relationship matrices (GRM) using both low and high density SNP panels, and computed the missing heritability through comparison to pedigree (NRM). Overall, the low density indicine panel performs very well in characterising the Brahman population. We found that estimation of missing heritability was broadly similar for both panels across the ten phenotypes. This implies similar amounts of genetic variation relevant to those phenotypes have been captured. The phenotypes with the lowest missing heritability (coat type and sheath score in Tropical Composites) possess an architecture that can be characterised simply. That is, they are dominated by genes of large effect.