- Of the top 10 journals shown in Figure 1, XY amount do not require this in their guide for authors.
Most scientific journals require depositing raw NGS data (preferably in FASTQ format) pertinent to reported research in public repositories such as NCBI (SRA or GEO archives) as a precondition of acceptance of manuscript for publications. This practice, however, does not seem to be sufficient to always safeguard reproducibility of the reported research. It seems that quite many datasets deposited in those archives are in fact results of the whole sequencing runs with no sufficient information/metadata provided to separate the individual amplicons, e.g. by reporting fully demultiplexed data or indexes/index combinations. Further, as only the raw output of sequencing experiments are available in the repositories in most cases, it may be difficult to reconstitute from such data the exact datasets used for the reported statistics and illustrations. This is because it needs precise reporting of all applied quality filters and processing steps (e.g. chimaera removal, rarefied datasets etc), including versions of the various software packages, not always fulfilled in every single paper. Therefore, we advocate reporting final sequencing results used for generating statistical tests with all relevant metadata along with each paper – either as a supplementary information accompanying the paper or deposited in public repository such as Dryad or similar.