Skip to main content
Fig. 1 | Microbiome

Fig. 1

From: A framework for assessing 16S rRNA marker-gene survey data analysis methods using mixtures.

Fig. 1

Assessment framework. a Count tables evaluated by the assessment framework are generated from the assessment dataset using marker-gene survey bioinformatic pipelines. Count table rows are features identified by the bioinformatic pipeline and columns are samples, four PCR replicates (labeled A–D) were sampled for PRE and POST and titrations, to simplify the diagram only three titrations are shown(labeled T1–T3). b The pictorial depiction and description of the seven feature types used in the assessment framework. c Qualitative and quantitative assessment metrics used in the assessment framework. Artifactual feature proportion (AFP) and Sparsity (SPAR) were used in the qualitative assessment. The artifactual feature proportion metric (AFP) is a qualitative assessment of feature presence/absence based on unmixed-specific or titration-specific artifactual features. Sparsity (SPAR) is a qualitative assessment of the proportion of observed features in each sample relative to the total observed features. For the quantitative assessment relative (Rel) and differential abundance (Diff), bias and variance metrics were used. For the quantitative assessment metrics, the relative abundance (Rel) and differential abundance (Diff) plots of example features are used to describe how the bias and variance metrics are calculated. For the relative abundance (Rel) metrics the error rate (|Obs-Exp|/Exp) is calculated for individual titrations and the bias (median(error)) and variance (RCOV) metrics are summaries the error rates by feature. For the differential abundance (Diff) plot points represent the log fold-change between two titrations, point text indicates the titrations compared. A linear model is fit to the data and the model fit information is used for the differential abundance bias (|1−slope|) and variance metrics (R2). Each feature type in b is labeled with the assessments shown in c in which they are employed

Back to article page