CSV Data Formats¶
synfit accepts dose-response data in several CSV formats depending on the experiment type.
Single-Drug: Tidy Format¶
One row per observation. Three required columns:
concentration,y,replicate
100.0,0.03,A
100.0,0.04,B
33.3,0.08,A
33.3,0.09,B
11.1,0.25,A
11.1,0.28,B
0.0,1.01,A
0.0,0.98,B
Columns are recognised by name: concentration (or conc, dose), y (or a name starting with response, viability, effect, inhibit…, activ…) and replicate (or rep, plate, well, group). Other names fall back to the first, second and third column. See how columns are detected.
Single-Drug: Wide Format¶
Concentrations in the first column, replicates in subsequent columns:
concentration,A,B,C
100.0,0.03,0.04,0.035
33.3,0.08,0.09,0.085
11.1,0.25,0.28,0.26
3.7,0.55,0.58,0.54
1.2,0.82,0.79,0.81
0.0,1.01,0.98,1.02
- N rows (concentrations) × R columns (replicates)
- The first column is treated as concentrations by default (configurable)
- Column names (A, B, C) become replicate labels
- Internally melted into tidy format for fitting
Matrix: Replicate CSVs¶
For drug combination experiments, each replicate is a separate CSV file containing a numeric matrix with no headers:
1.02,0.98,0.85,0.45,0.12
0.97,0.94,0.80,0.41,0.10
0.88,0.85,0.72,0.35,0.08
0.52,0.48,0.38,0.18,0.04
0.15,0.12,0.09,0.05,0.02
- Rows correspond to vertical drug concentrations (top to bottom)
- Columns correspond to horizontal drug concentrations (left to right)
- Drug concentrations are provided separately during upload (not in the CSV)
- Upload one CSV per replicate (e.g., rep1.csv, rep2.csv, rep3.csv)
Metadata During Upload¶
In addition to the CSV data, the upload screen collects:
- Drug name — label for the compound (single-drug) or both compounds (matrix)
- Concentration unit — nM, µM, or mM (per drug for matrices)
- Direction — inhibition or activation (per drug for matrices)
For matrix uploads, you also enter the concentration of each row and column, by hand or with the serial-dilution generator.
Common Issues¶
Numeric values only
Response values must be numeric. Text entries (e.g., "N/A", "failed") will be dropped as NaN during parsing. If too many values are dropped, the fit may fail or produce unreliable results.
Separator
Use commas or whitespace as separators. Semicolons, tabs, and other separators may not be parsed correctly for matrix files.
Zero-concentration control
Including a zero-concentration point (no drug) helps anchor the Top parameter. For matrix files, the first row and first column should correspond to zero concentration of one drug.