Figure 1.
(A) Overlap between three database systems showing their linking Python scripts and the central item, ‘Strain ID’, which is common to all three. (B) Modified data structure of the ItemTracker SMS, displaying types and hierarchies of data items. Item levels are indicated by rectangles containing important item properties under the Item type designation that was devised for the modified data structure. All item properties are presented in Tables S3 and S4. Arrows indicate hierarchical parent-child relationships, except that ‘StrainID’ is inherited by all decendents of ‘Bacteria’. Each item has a unique ‘ItemID’. Each of the item levels other than ‘Bacteria’ also includes a second, user-understandable, unique designation, ‘ItemName’. (C) Hierarchy of locations in the SMS, including examples of location designations in the lower line.
Table 1.
Number of Items in sample management system.
Figure 2.
Overview of five discrete modules within the pipeline with indications of their throughput and principles.
This figure includes photographs of the four liquid handler systems, LHS1-4, which are shown in greater detail in Fig. S1. Still other details on LHS1 are in Fig. 3A and File S2. Additional details on Module E are in Fig. 5. GUIs used in Modules A and D/E are in Figs. S2 and S3, respectively.
Table 2.
Time and costing of pipeline for processing 96 strains.
Figure 3.
Automated sub-cultivation with LHS1.
(A) Schematic layout of the equipment within the LHS1 cabinet. Positions A and B represent stack units where racks of 2D tubes and deep well plates are stored prior to operations. Position C designates the 5-position, XY-moveable stage for the 96-tip pipettor. Position D represents a rest position used for regripping of racks of 2D tubes before and after capping/decapping in position G. Positions E and F represent 2-D and 1-D bar-code readers. H is a SCARA robotic arm for moving between these positions. Further details of how LHS1 was used are presented in File S2, Figs. S6/S8 for sterile dispensing of media into 2D tubes, and S7/S9 for aseptic sub-cultivation of bacteria into 2D tubes and deep well plates. (B) Improvements in pipetting parameters (right) which eliminated aerosol formation that led to cross-contamination between cultures (left).
Figure 4.
Yields of DNA after automated extraction and normalisation in Module B.
(A) Notched box and whiskers plots of DNA yields after automated DNA extraction from S. enterica or L. monocytogenes. The notch indicates the 95% confidence estimate of the median value (central line), which splits the boxes in the second and third quartiles of the data. The first and fourth quartiles are indicated by the external horizontal lines with outliers shown by single circles. The number of samples that has been summarized is indicated in parentheses under the bacterial designations on the X axis. The left scale reflects the DNA concentration and the right scale is the total yield within the 150 µl elution volume. (B) DNA concentrations for one rack of 96 tubes of S. enterica after elution as in part (A) (empty squares), and after automated normalisation (solid circles) to 3.3 ng/µl.
Figure 5.
Overview of robotic sequencing in Module E.
Top. 1-D and 2-D bar-codes of DNA tubes and 384-well plates are scanned, and information for genes sequencing, including orientation, is collated from the Cherrypicking database for the scanned DNA tubes. Bottom. Left: a pipetting layout which optimises pipetting efficiency is calculated by Script E7 for one to two 384-well plates. Color-coded arrows indicate different genes for each DNA which are to be sequenced. Right: LHS4 dispenses primer mixes from 2 ml tubes in a cooled thermoshaker to a 384-well plate, followed by DNA.