Here is where I think you are confused.
It is true that one means of determining whether a particular DNA sequence if “functional” is by determining the degree to which it is conserved (i.e. subject to selection). Note that, strictly speaking, this is not a means of defining function, but rather of measuring it.
You would be correct if your position is that this is a method that relies on accepting evolution and, in particular, the concept of common ancestry between different species.
Dan Graur may well favour this as a means of determining whether a sequence is functional. However, this is not germane to the particular paper we are discussing. He does not compare sequence conservation in the human line to that of other lineages.
Here is what his paper did: He starts from the premise that, the higher the number of detrimental mutations occurring in a population, the greater the number of offspring must be produced in order to prevent extinction due to the poor fitness of offspring affected by detrimental mutations. This is non-controversial.
The average number of detrimental mutations per offspring can be estimated by knowing 1) the overall mutation rate; 2) the percentage of mutations that are detrimental if they occur in a sequence that is functional and 3) the percentage of the genome that is functional.
So, to use some hypothetical numbers to illustrate:
If the mutation rate is such that 100 mutations occur a generation, 10% of these are detrimental (in functional sequences) and 50% of the genome is functional, this will produce 5 detrimental mutations per generation. But if only 10% of the genome is functional, that number goes down to 1.
Is that clear? If not, let us know and someone I’m sure can explain further.
Graur then went on the calculate the numbers for a range of mutation rates and ratios of detrimental to neutral mutations as applied to various proportions of the genome that could be functional.
He found that at any levels of function >25%, unrealistic numbers of offspring per couple would be required to prevent the population from going extinct from the genetic load of deleterious mutations.
Now, the point is that this has nothing to do with evolutionary concepts such as common ancestry. As I said, it only requires acceptance of the existence of genetic inheritance and mutations, which applies to every single creationist I can think of. Graur’s argument is independent of evolution. If he was wrong, evolution could still be true. And if he is correct, evolution could still be wrong