In one of the last Newsletter issues (LINK), we wrote about the uncertainties which sometimes exist when trying to find the best normalization approach for a given study. In this context, the need to be transparent about normalization or randomization steps and provide sufficiently detailed information is nicely demonstrated by the following example:
One of the most influential papers about the role of the gut microbiota in obesity has been published in Nature in 2006 by Turnbaugh and colleagues. It has already been cited more than 8000 times. The reason for its importance is that this was the first study to show that transplanting the microbiota from obese mice into lean mice made the mice fatter than mice getting transplants from lean mice.
The key figure is shown below, demonstrating that, in germ-free mice transplanted with gut microbiota from obese donors (n = 9 recipients), body fat increased by 47% compared to a 27% increase when donors were lean (n = 10 recipients).
This study outcome (Figure 1) looks indeed quite impressive and apparently merited a publication in Nature.
Unfortunately, however, the paper does not provide an explanation how these 19 recipient mice were initially randomized and allocated to the two different groups and whether or not body fat/weight was considered as a normalization denominator.
Why is this relevant?
It is important to note that this figure does not show body fat percentage but rather the percentage increase in fat in relation to the initial body fat.
Although the paper does not state the starting body fat of the mice, from the information provided it is possible to count back (nicely done by Matthew Dalby) to end up with the following situation:Mice getting transplants from obese mice had 2.76 grams of fat at the start.Mice getting transplants from lean mice had 3.19 grams of fat at the start.The final body fat of mice getting transplants from obese mice was 4.07 gramsThe final body fat of mice getting transplants from lean mice was 4.05 gramsThis is summarized in Figure 2, which certainly does not look as impressive as Figure 1.
Although the mice getting a fecal transplant from the obese mice gained a little more fat, they ended up with the same body fat at the end of the study as the mice getting transplants from lean mice. As mice at this age would typically weigh about 25 grams, at 4 grams of body fat both groups of mice would be considered lean and therefore neither group of mice were obese after the transplantation.
Are we too critical? We would certainly have less reasons to review this example if more information was provided in this publication about the randomization process and whether the normalization was pre-specified.