Fig. 1: Experimental design and species richness for community sequence data generated with varying levels of workflow autonomy.

a The typical steps (e.g. DNA extraction, PCR amplification, and DNA sequencing) involved in generating soil microbial community sequence data were evaluated as sources of variability. The bars above the steps denote the processes performed by secondary labs to generate each of the three libraries (Ext/PCR/Seq, PCR/Seq, and Seq) that were pooled prior to sequencing on the MinION platform (ONT). b Species richness for two soil sites (n = 12) generated in each sequencing run (Lab1, Lab2.a, etc.) and within each of the three pooled libraries. Data were rarefied to 12,000 reads prior to calculating richness estimates. Each box depicts the interquartile range (IQR), where the bottom of the bar is the 1st quartile (Q1), the middle bar is the 2nd quartile or median, and the top of the bar is the 3rd quartile (Q3). Whiskers are calculated as 1.5 × IQR above Q3 or 1.5 × IQR below Q1 and points outside this range are outliers. c The relationship between the original number of sequence reads and observed species richness for each sample. Left: Michaelis–Menten plot where points are coloured by sequence run and equations were fit separately for each site (ARDEC, solid line; Pendleton, dashed line). Right: Lineweaver–Burk plot where points and lines are coloured by laboratory and equations were fit separately for each sequence run. All images in this figure are original artwork by the authors.