One thing we have learned from the past 20 years of mtDNA genetics in patients is that you should pay attention to both the absolute mtDNA copy number and the proportion mutant.  Both numbers give you useful information.  Back in the 90s Patrick Chinnery and I came up with the "Maintenance of Wild-Type" hypothesis that cells respond to a shortfall in mitochondrial function by ramping up mitochondrial biogenesis to compensate.  When the mitochondrial dysfunction is due to an mtDNA mutation, this feedback leads to increased mtDNA replication, which brings the absolute number of wild-type mtDNA up to approximately normal numbers, maintaining mitochondrial function.  A side effect of this increased biogenesis is that the number of mutated mtDNA molecules would also increase, assuming both wild-type and mutant mtDNA had equal probabilities of replication.

You don't give details of what you mean by "treatment", but many drugs and other chemicals adversely affect mitochondrial function and I would expect to see general mtDNA proliferation in those cases, as the cell attempts to compensate for the lowered mitochondrial function.  In that case, the deletion decline in absolute numbers is surprising to me. 

My main answer is that you should be reporting two measures, total copy number and either mutation copy number or proportion mutation.  But two separate numbers should be reported for a full picture.

I need to know more. What cells? Or what animal and what organ? What sampling method? What is the treatment? 

Let me restate this and correct it if I am incorret. You are saying you are using PCR to look at mtDNA copy number using some primer(s). In response to your drug (or some other intervention?) you see the mtDNA copy number increasing.

But you also see that mtDNA with deletions is going down. The way it is stated, you have seen absolute numbers drop. But since mtDNA copy number overall is increasing, you might be seeing the percentage drop.

Generally speaking, I would suggest the question indicates there may be a lack of full understanding of the meaning of what you are looking at. Because copy number itself is meaningless relative to a treatment in any case I can think of. The mtDNA copy number that matters is the deletion copy number. Because it's the deletion that causes pathology in aged, or damaged cells. You need both. There is no normalization that makes sense here unless you are doing something unusual.

Have you talked to your PI about it?

You need to examine both total mtDNA copy number and what proportion of the mtDNA population is intact and what proportion has the deletion.  You need a way to measure both types of mtDNA separately from each other.

One thing we have learned from the past 20 years of mtDNA genetics in patients is that you should pay attention to both the absolute mtDNA copy number and the proportion mutant.  Both numbers give you useful information.  Back in the 90s Patrick Chinnery and I came up with the "Maintenance of Wild-Type" hypothesis that cells respond to a shortfall in mitochondrial function by ramping up mitochondrial biogenesis to compensate.  When the mitochondrial dysfunction is due to an mtDNA mutation, this feedback leads to increased mtDNA replication, which brings the absolute number of wild-type mtDNA up to approximately normal numbers, maintaining mitochondrial function.  A side effect of this increased biogenesis is that the number of mutated mtDNA molecules would also increase, assuming both wild-type and mutant mtDNA had equal probabilities of replication.

You don't give details of what you mean by "treatment", but many drugs and other chemicals adversely affect mitochondrial function and I would expect to see general mtDNA proliferation in those cases, as the cell attempts to compensate for the lowered mitochondrial function.  In that case, the deletion decline in absolute numbers is surprising to me. 

My main answer is that you should be reporting two measures, total copy number and either mutation copy number or proportion mutation.  But two separate numbers should be reported for a full picture.

I suggest to check mtDNA copy number by qPCR using 2 different regions:

- the D-loop is typically a non-deleted region, if you lose it, the molecule will not be maintained because it cannot be replicated anymore. When compared to a single copy nuclear gene, this should give you the total mtDNA copy number (all molecules, deleted and non deleted)

- one other region from the major arc (between OH and OL), such as ND4, which can be subjected to deletions. When normalized to the single copy number nuclear gene, you can afterwards compare this to the copy number determined with the D-Loop primers, and have an estimate of the proportion of molecule harboring ND4 deletions.