Since heat is the kinetic energy possessed by atoms [...]
That is not exact. If you work in the approximation of free particles (e.g. in an ideal gas) then heat is at all related to the kinetic energy, but beyond that approximation potential energy also plays a role. If U is internal energy
U = Sum_j p_j E_j
and E_j is energy of level j and p_j its probability, then heat is often defined as
dQ = Sum_j (dp_j) E_j
Notice that E_j is not restricted to be only a kinetic energy.