International project puts forward universal stool study protocols

The project aims to establish an international consensus that those working on the intestinal microbiota can agree upon to ensure findings are consistent and more easily interpreted.

The team, headed up by the French National Institute for Agricultural Research (INRA) and the Institute of Genomics (CEA) based near Paris, propose a suite of standardised procedures for processing samples, from the collection to the bioinformatic analysis of the data.

“Knowing the human intestinal microbiota and assessing the impact on health requires standardising the metagenomic methods used,”​ said the team.

“While the international scientific community has invested heavily on the subject, there has been no standard protocol. ​

“We propose an efficient, transferable and automated for DNA extraction, which will contribute, in the field of the human microbiota and beyond, to optimise data quality and their comparison.”​

Lower IBD gut bug diversity ​

Current commercial DNA extraction kits or laboratory protocols exhibit a wide inconsistency that the researchers believe affects results, more so than biological factors such as differences between individuals or other technical procedures such as DNA preservation.

Studies that link specific microbiota to diabetes, inflammatory bowel disease (IBD) or colorectal cancer ​have almost always used their own methodology and a distinct sample group meaning comparisons or meta-analyses are limited in their interpretability.

For example, healthy Americans from the Human Microbiome Project study show lower taxonomic diversity in their stool than patients with IBD from a European study, although it is known that IBD patients worldwide have reduced taxonomic diversity.

According to the project’s leaders, this illustrates the difficulties in disentangling biological from technical variation when comparing across multiple studies.

The team’s comments are evidenced in a study, in which 21 DNA extraction protocols were compared in 11 countries located in the northern hemisphere across Europe, America and Asia.

Of the 366 bacterial species tested, 90 (25%)–made up of mostly Gram-positive bacteria (around 37% of total microbiota bacteria)—were affected by the extraction protocol and were thought to be under-represented.

The team believed that the bacteria’s cell wall contributed to these figures, as it is considered more resistant to the mechanical forces mobilised during the extraction process.

They then assessed the transferability and accuracy of the most efficient DNA extraction protocols using a synthetic microbial community consisting of strains representing 10 bacterial species.

These evaluations formed the basis of a set of standard protocols (Protocol Q), put forward by the INRA, the CEA and their partners for the treatment of human stool specimens, from sampling to bioinformatic analysis, including extraction and sequencing of DNA.

‘Protocol Q seems the best’​​

“Although for particular applications some of the tests are more important than others, reproducibility across labs is more important than in an in-depth study in one location,”​ the study said, “Protocol Q seems to be the best overall and should suit most applications.​

“We further tested the quantification accuracy of the best-performing protocols by using a mock community and showed that protocol Q had a median absolute quantification error of less than 0.5×.”​

The team acknowledge that while they have only tested this protocol on stool, it may also work well with other samples.

“However, we caution that additional considerations may apply, such as that of kit contamination^, ​which may differ between the protocols investigated here and would, for example, have a high impact on samples with low biomass.”​

Protocol Q, together with standard methods for sample collection and library preparation, can be found on the IHMS website​​.​

Source: Nature Biotechnology​

Published online ahead of print: doi:10.1038/nbt.3960​​

“Towards standards for human fecal sample processing in metagenomic studies.”​

Authors: Paul Costea et al.