Impact of technical sources of variation on the hand microbiome dynamics of healthcare workers

Abstract

We assessed the dynamics of hand microbial community structure of 34 healthcare workers from a single surgical intensive care unit over a short (3 week) time period, whilst taking into account the technical sources of variability introduced by specimen collection, DNA extraction, and sequencing. Sample collection took place at 3 different time points. Only the sampling collection method appeared to have a significant impact on the observed hand microbial community structure among the healthcare workers. Analysis of samples collected using glove-juice showed a slightly more similar microbial composition within individual hand samples over time than between the hands of different individuals over time. This was not true for samples collected using a swab, where samples from a single individual were no more similar to each other over time than those among other individuals over time, suggesting they were essentially independent. DNA extraction techniques (lysozyme only versus enzyme cocktail) and sequencing (replicate set 1 versus 2) using Ion Torrent Personal Genome Machine, were not influential to the microbial community structures. Glove-juice sample collection may likely be the method of choice in hand hygiene studies in the healthcare setting.

Figure 1. Study Design Showing Levels of Comparisons of Hand Microbiota Samples Sent for Sequencing.

Level A shows the comparison of within versus between HCWs (n₁ = 34, n₂ = 34, n₃ = 34); level B shows the comparison of sampling collection methods (n_SW = 102, n_GJ = 102); level C shows the comparison of sequencing replicates (n₁ = 30, n₂ = 30); and, level D shows the comparison of DNA extraction methods (n_C = 10, n_L = 10). Samples obtained from 34 Healthcare Workers at the University of Michigan Surgical Intensive Care Unit, July 5–28, 2011.

Figure 2. Within and Between Weighted UniFrac Distances of the Hand Microbiota.

Stratification by Sampling Collection Method (Panel A: Glove-Juice and Swab), DNA Extraction Method (Panel B: Lysozyme and Cocktail), Sequencing Replicates (Panel C: Set #1 and Set #2), Healthcare Workers (Panel D: Within and Between), and Healthcare Workers by Sampling Collection Method (Panel E: Within and Between). Samples obtained from 34 Healthcare Workers at the University of Michigan Surgical Intensive Care Unit, July 5–28, 2011.

Figure 3. First Principal Components Scatterplots of the Principal Coordinate Analysis (weighted UniFrac) of the Hand Microbiota.

Stratification by Sampling Collection Method (Panel A: Glove-Juice (x-axis) and Swab (y-axis)), DNA Extraction Method (Panel B: Lysozyme (x-axis) and Cocktail (y-axis)), and Sequencing Replicates (Panel C: Set #1 (x-axis) and Set #2 (y-axis)). Samples obtained from 34 Healthcare Workers at the University of Michigan Surgical Intensive Care Unit, July 5–28, 2011.

Figure 4. Relative Abundances of the Top 80% Most Abundant Taxa Detected per Sampling Method.

Sampling methods (Panel A: Glove-Juice; Panel B: Swab) of the Hand Microbiota are obtained from 34 Healthcare Workers at the University of Michigan Surgical Intensive Care Unit, July 5–28, 2011.