SOURCE DATA: CUMA-OHC14-ATC THREE-SENSOR COMMUNITY

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

Folder / archive
03_Cuma_OHC14_aTc

Dataset description
This archive contains the experimental plate-reader time courses for the three-member Escherichia coli sensor community used to detect cuminic acid (Cuma), N-(3-hydroxytetradecanoyl)-DL-homoserine lactone (OHC14), and anhydrotetracycline (aTc). The Cuma-responsive strain reports through YFP, the OHC14-responsive strain reports through CFP, and the aTc-responsive strain reports through mCherry.

The three workbooks collectively contain 777 co-culture time-course samples spanning 281 unique Cuma-OHC14-aTc input combinations. The first workbook also contains one blank well. In the third workbook, 84 rows are labeled "bad wells," one row is labeled "atc is off," and one row is labeled "outlier." These explicitly flagged rows are retained in the original workbook for transparency but are not included in the 777-sample dataset total reported in the manuscript.

Files
1. 2023-07-14 corrected cuma atc ohc14 mixed plate full timecourse.xlsx
   - 95 co-culture samples
   - 64 unique Cuma-OHC14-aTc input combinations
   - 1 blank well
   - 232 recorded time points, spanning 0-69,298.3 s

2. 20240224 cuma ohc14 atc.xlsx
   - 384 co-culture samples
   - 128 unique Cuma-OHC14-aTc input combinations
   - 86 recorded time points, spanning 0-86,099 s

3. 20240229 cuma ohc14 atc.xlsx
   - 298 co-culture samples
   - 96 unique Cuma-OHC14-aTc input combinations
   - 84 rows labeled "bad wells"
   - 1 row labeled "atc is off"
   - 1 row labeled "outlier"
   - 88 recorded time points, spanning 0-88,098 s

Some input combinations were measured in more than one workbook; therefore, the sum of the workbook-level unique-combination counts exceeds the 281 unique combinations in the combined dataset.

Workbook organization
Each workbook contains four worksheets:
- OD: optical-density time courses
- yfp or YFP: yellow-fluorescence time courses
- cfp or CFP: cyan-fluorescence time courses
- mCherry or mcherry: red-fluorescence time courses

The worksheets use the same sample-row ordering within each workbook. Each data row represents one microwell sample. The first five columns contain:
1. Sensor: sample status ("both" for the three-strain co-culture, "blank" for the media-only well, or an explicit exclusion label in the third workbook)
2. cuma: Cuma input concentration
3. ohc14: OHC14 input concentration
4. atc: aTc input concentration
5. Sample or plate identifier; in the original workbooks this column is labeled "Time [s]"

Columns 6 onward contain measurements at successive time points. Their column headers are elapsed time in seconds. OD values are plate-reader optical-density measurements; YFP, CFP, and mCherry values are plate-reader fluorescence measurements. Input-concentration values are preserved exactly as recorded in the experimental workbooks.

Figure mapping
These files provide the experimental source measurements for:
- Main Figure 4a-b
- Supplementary Figure 12c
- Supplementary Figure 13c
- Cuma-OHC14-aTc entries in Supplementary Table 3

Crosstalk values, fitted dose-response curves, augmented datasets, latent representations, and model predictions shown in these figures 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. Rows explicitly labeled "bad wells," "atc is off," or "outlier" in the third workbook were excluded from the reported 777-sample co-culture dataset and downstream analyses. As described in the Methods, samples with zero inducer concentration or concentrations outside the 1st-99th percentile of the corresponding sensor dynamic range were excluded from specified machine-learning analyses. The unfiltered source measurements, including zero-input conditions used for baseline, crosstalk, and dose-response analyses, are retained in these files.

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
