We are not very interested in recording "transgender crimes". What we are most interested in is clearly and separately recording the crimes committed by biologically male people and the crimes committed by biologically female people so that they can be compared. Whether these people are transgender or not doesn't make much difference to their offending rates.
crime reporting & incarceration is not reflective of offending given most sex crimes aren't reported.
Ah, you have a common misunderstanding. The truth is that crime reporting and incarceration are highly reflective of differences between the biological sexes, even despite the fact that most of these crimes are not reported.
There is no evidence of sex crimes committed by biologically female people being disproprotionately unreported - these crimes may be under-reported, but only at about the same rate as for men. So the under-reporting doesn't change anything. There aren't squillions of women secretly committing sex crimes getting away with it and distorting the figures. So sex crime is still overwhelmingly a crime committed by biological males.
Nor is there evidence that biologically male people with a trans identity commit sex crimes at a lower rate than other men do. If you hope to find such evidence then yes you could record sex crimes committed by biologically male people with a trans identity separately from men without, and then compare the two.
What you must not do is to count even a small number of high-risk individuals (biological males) in a very low-risk population (biological females) because this massively distorts the resulting figures. So if you lump together sex crimes commited by biologically male people with a trans identity in the same category as sex crimes commited by biologically female people it creates an inconsisent mess which makes women appear disproportionately more dangerous and distorts the focus of resourcing and policing away from protecting women.