A key part of this evaluation is validation, where metaMATE independently checks the ASVs to determine if any are clearly valid or invalid, using this control group to evaluate the effect of different abundance calculation and threshold strategies. This means that it bases its filtering on ASV frequencies, but it calculates frequencies based on multiple strategies, and then evaluates many different threshold values to generate data from which the ideal filtering strategy can be determined. This pipeline is centred around the tool metaMATE, which performs so-called “Multidimensional Abundance Threshold Evaluation”. This pipeline is not exclusively for studies exploring ASV patterns, the more conservative filtering of ASVs will also improve OTU results as well. In this extension we will look at an alternative pipeline that builds upon the standard pipeline presented elsewhere to more stringently filter ASVs. Secondly, we must be confident that our error filtering has not only removed highly divergent errors that might have caused false OTUs, but also removed minor errors that would have otherwise been absorbed into OTUs. Firstly, our sampling must have been deep enough to capture not just sufficient species richness but sufficient haplotype richness to answer our research questions.
When we wish to use ASVs for ecological analyses at the level of populations, we may need greater confidence in the accuracy of our ASVs than compared with cases where we intend to cluster these ASVs to species.
0 kommentar(er)
