SOURCE DATA: HOSPITAL SINK-WATER SENSOR COMMUNITY

Paper
Engineering microbial consortia for distributed signal processing
Nature Communications manuscript NCOMMS-26-017824B

Folder / archive
06_Hospital_sink_water

Dataset description
This archive contains the experimental plate-reader measurements used to evaluate inference of spiked anhydrotetracycline (aTc) and vanillic acid (Van) in pooled hospital sink-water. The sensor community contained an aTc-responsive strain reporting through mCherry and a Van-responsive strain reporting through YFP.

Sink P-trap filtrates from five untreated control sinks collected during week 12 were pooled at equal volume and mixed with LB medium at a 3:1 sink-water-to-LB ratio, as described in the Methods. The analytes were added at defined concentrations before time-resolved measurements of optical density, mCherry, and YFP.

The two machine-learning experiment workbooks collectively contain 189 valid co-culture time-course samples spanning 93 unique aTc-Van input combinations. These are the dataset totals reported in the manuscript.

Files
1. 2024-12-08 atc-mch van-yfp sink ml experiment.xlsx
   - 96 valid two-strain co-culture samples
   - 64 unique aTc-Van input combinations
   - 275 recorded time points
   - elapsed-time range: 0-82,208.1 s

2. 2024-12-10 atc-mch van-yfp sink ml experiment.xlsx
   - 93 valid two-strain co-culture samples
   - 64 unique aTc-Van input combinations
   - 271 recorded time points
   - elapsed-time range: 0-81,008.1 s
   - also contains two wells explicitly labeled "leaked" and one blank well; these three rows are retained in the source workbook but excluded from the reported 189-sample dataset total

Thirty-five input combinations occur in both machine-learning workbooks. Consequently, the combined dataset contains 93 unique combinations rather than the sum of the workbook-level counts.

3. 2024-12-04 hospital sink week 12 atc van crosstalk.xlsx
   - dedicated control experiment comparing sensor responses in pooled sink-water and LB
   - contains aTc-mCherry monocultures, Van-YFP monocultures, the two-strain co-culture, blank controls, and unused plate positions
   - 81 wells contain recorded measurements:
     24 aTc-mCherry wells
     24 Van-YFP wells
     19 co-culture wells
     14 blank wells
   - 5 recorded time points spanning 0-89,700 s
   - supports Supplementary Figure 18 and the aTc-Van crosstalk entries in Supplementary Table 3
   - Supplementary Table 3 reports n = 2 replicate wells per condition for the displayed crosstalk estimates

Workbook organization
The two machine-learning workbooks each contain three worksheets:
- OD: optical-density time courses
- mCherry: red-fluorescence time courses
- YFP: yellow-fluorescence time courses

The crosstalk workbook contains:
- OD
- mcherry
- yfp

Within each workbook, the worksheets use identical sample-row ordering and metadata.

Machine-learning workbook columns
The first four columns contain:
1. Sensor: sample type ("both" for the two-strain co-culture; additional labels identify leaked or blank wells where applicable)
2. atc: aTc input concentration
3. van: Van input concentration
4. Time [s]: microwell identifier

Columns 5 onward contain measurements at successive time points. Column headers are elapsed time in seconds.

Crosstalk workbook columns
The first five columns contain:
1. Sensor
2. media: "sink" or "lb"
3. atc
4. van
5. Time [s]: microwell identifier

Columns 6 onward contain measurements at the five recorded time points.

Input-concentration values are preserved exactly as recorded in the experimental workbooks and are interpreted by the archived analysis code. OD values are plate-reader optical-density measurements; mCherry and YFP values are plate-reader fluorescence measurements.

Figure mapping
- Main Figure 6a:
  Experimental-workflow schematic; no numerical source data are required for the schematic itself.

- Main Figure 6b:
  2024-12-08 atc-mch van-yfp sink ml experiment.xlsx
  2024-12-10 atc-mch van-yfp sink ml experiment.xlsx

- Supplementary Figure 18:
  2024-12-04 hospital sink week 12 atc van crosstalk.xlsx

- Supplementary Table 3, aTc-Van entries:
  2024-12-04 hospital sink week 12 atc van crosstalk.xlsx

The fitted dose responses, augmented simulations, latent representations, and predicted concentrations shown in the manuscript are computationally derived from these experimental measurements and can be regenerated using the archived analysis code, fitted parameters, and model files.

Data processing and exclusions
The complete experimental workbooks are supplied here. The two rows labeled "leaked" and the blank row in the 2024-12-10 workbook are retained for transparency but are not included in the 189 valid co-culture time courses reported in the manuscript.

As described in the Methods, specified machine-learning analyses excluded samples with zero inducer concentration or concentrations outside the 1st-99th percentile of the corresponding sensor dynamic range. The unfiltered source measurements, including zero-input and control conditions used for baseline and crosstalk analyses, are retained in these workbooks.

Plate-reader calibration
Cross-instrument conversion tables used during preprocessing, where applicable, are supplied separately in the 08_Plate_reader_calibration Source Data archive.

Code and reproducibility
Analysis code is available at:
https://github.com/youlab/multiplexed_sensing

The analyses were performed using Python 3.12.7.

Contact
Lingchong You
Department of Biomedical Engineering, Duke University
Email: you@duke.edu
