Introduction
This is a (more) critical review of the recently published article ...
I am not criticising the limits of the experimental design. I recognise that comprehensive studies are expensive and often difficult to conduct. The problem I have with this article is of a different nature. Based on my re-analysis of their data, I can only agree with the authors on two aspects of the study: (a) Different tissue samples of an apple harbour different microbiota communities (at least most do), and (b) I also agree that when we eat an apple, we are consuming bacteria. For the rest, I have difficulty reproducing, understanding and trusting it.
Bacteria Cells per Apple - In the abstract, the authors first claim that there are 108 gene copies of 16S rRNA per 1g of apple. A few sentences later the authors state that we eat about 10^8 bacteria cells per apple. This can only be true if an apple is 1g on average and gene copy number is equal to bacteria cell counts. We know that both assumptions are not true. My estimate of bacterial cells per apple based on the provided data is around 15 million, and there is no difference in the amount of bacteria between treatments (organic-fresh vs. non-organic-supermarket-not-fresh).
Organic vs. Non-organic - According to the authors, one of the main aims of the study was to provide insights of the impact of organic and conventional management practices on apple fruit-associated bacteria. In fact, the study compares unwashed supermarket apples with freshly picked organic apples. Not discussing this aspect, pooling tissues and reducing the treatment to organic versus conventional grown apples is unfair. I suspect that the post-harvest treatment of non-organic apples has an influence on the apple fruit-associated bacteria, at least on the outside.
Hot Spots - The authors declare fruit pulp and seeds as bacterial hot spots. This makes little sense. Why would bacteria diversity be higher inside an apple than outside? The authors acknowledge the co-amplification of chloroplast and mitochondria rRNA but they do not quantify it. Based on their own data, there is clear evidence that co-amplification can be substantial (in tissue samples with low bacterial biomass), and we cannot just ignore it. Therefore, I do not believe that fruit pulp and seeds are bacterial hot spots.
Pooling Data - The authors pooled biological replicates and ignored intra-sample variation entirely. This is a fundamental flaw, since the variation between replicates is substantial. The comparison between tissues is also questionable because the weight of the raw material differs and they have used different extraction methods. The authors combined tissues to compare treatments. This is again problematic because the tissue samples derived from the same apple and are therefore not independent. As a result, the relevant health-related claims of the paper are suspicious and important information got lost and mixing exterior and interior tissue samples.
Unsupported Claims - The introduction and discussion are all about food safety, but the study provides little or no evidence since no food-borne pathogens were clearly identified. The authors declare Enterbacteriales a signature taxa for non-organic apples, but this is only true for two tissue samples. The authors also claim that Escherichia-Shigella is specific to non-organic apple. While most Escherichia-Shigella are harmless, I could only find credible amounts on peel samples. The authors further allege that Lactobacillus belongs to the core taxa of organic apples. My analysis shows a different picture where Lactobacillus is rare and most abundant on peel samples of conventional (supermarket) apples. The interpretation of the data is fictional and does not provide any health relevant information.
Errors - There are serious mistakes in the data analysis. The authors remove about 99% of the bacterial data, base their argumentation only on rare taxa, focus solely on presence/absence instead of abundances, and ignore intra-sample variation. P-values for multiple pairwise Shannon diversity comparisons between treatments are uncorrected. The authors ignore important details throughout the paper.
Reproducibility - The authors leave the reader often with vague, conflicting or even missing descriptions. The disregard of reproducibility in the data analysis in combination with some questionable results (e.g., 92,365 OTUs, more outliers than data points) and the inconsistent writing does not encourage confidence in the study.
Experimental Design - Sample design does not contain any negative controls. The materials and methods do not show any precautions to reduce contamination. DNA extraction was heterogeneous (blender or mortar) and carried out by different students (see acknowledgement). Contamination of any kind (e.g., experimentator, extraction method, cross-contaminations, cross-talk) can not be excluded. The high level of variation between samples could show technical problems and should not be treated as biological relevant information.
Based on the re-analysis of the data provided by Wassermann et al. 2019, I cannot confirm any of their key findings. I have serious doubts about the data analysis, and I think they misinterpret their results.
One more thought - Is more better?¶
There is a mixed message in this paper and its media coverage. I do not understand why I have to eat the not so tasty core. Is it to get more or better bacteria? Is there a daily recommended dosage of bacteria or is the low abundance of Lactobacillus the reason?
The bacteria found in seeds are claimed to be mostly from the inside. Therefore, I would need to chew or grind the seeds to access it. Now, we know that apple seeds have a bitter taste and there is a reason for that. The seeds of fruits in members of the rose family contain amygdalin as part of the chemical defence. If amygdalin is released it degrades into hydrogen cyanide. While amygdalin is harmless, cyanide is not. The question is now, when does the benefits outweigh the risks.?
Agreed, you would have to chew hundred of seeds to experienced any adverse symptoms. Nevertheless, it's probably a good idea to remove the core before offering apples to children. I think it is safe to say that apple peel and flesh are healthy and pose no risks to your health. It is generally recommend that you wash your produce to avoid microorganisms capable of causing disease organic or not.
~