I want to make some comments on things in this thread.
First, I do not consider the Light data to be in any major conflict with the metabolomics data.
There are probably a number of hypotheses in which mitochondrial gene variants might increase risk of ME, but not be causal of ME. It might imply that they are part of the trigger process though, and so might be involved in severe relapses as well.
Risk from mitochondrial gene variants causing ME would indeed increase risk through the female line, because children inherit the mother's mitochondria. However if the gene is actually encoded in the nucleus, as some mitochondrial protein genes are, then it would be both the male and female lime that lead to inheritance. So it might be a confused mix.
As I see it the genetics establish risk. That is during extreme biological events the chance of something going wrong is higher. This does not mean that once the problem is established the primary driving force is the genetics. We need to know a lot more.
Rather than be alarmed that different researchers are pursuing different angles, I am encouraged. Once we have enough of the jigsaw pieces, and have looked at it long enough, the picture will begin to emerge. If everyone were working on the same tiny bit then we would only see a piece of the puzzle and might guess wrong about the big picture.
Discrepant data, conflicting data, opposing hypotheses etc. can be very good in science. In theory it makes scientists focus, analyze more carefully, and design better experiments.