CellFate · How it works
A lower signal is not an explanation. CellFate helps identify what changed in an experiment, which conclusions the measurements support and what is still missing.
Explore the learning example ↗Cell model
Fibroblasts or a cancer cell model? Specify the line, cell state and research question. This determines which comparisons make sense.
Human fibroblasts; in oncology, established human cell lines in 2D monoculture. Organoids and clinical samples are not validated scenarios.
Input measurements
CSV, Excel or a saved JSON project. Map columns and identify controls, doses, times and independent replicates. Processed fibroblast values have a separate mode.
Technical wells do not become independent experiments. Means, SD and SEM remain distinct quantities.
Data quality
CellFate checks table structure, missing values and comparison conditions. Paired analysis needs control and treatment matched within each independent replicate.
Without baseline cell counts, there is no GR. Without direct event observations, division and death rates cannot be calculated. Missing measurements are not invented.
Signal and cell count
Compare live-cell counts and energy signals separately. With suitable paired measurements, normalize the signal by cell count. This shows how much of a lower signal accompanies cell loss.
In the learning example, 50% of cells and 40% of ATP remain: ATP per cell is 80% of control. Actual analysis calculates ratios within independent replicates first.
Dose and time
Choose a condition, readout and time. Inspect observed values, controls and replicate spread. In oncology, suitable baseline data enables the GR growth-rate correction.
A line between measured points helps read the plot. It does not predict a later time. Flat cell counts alone cannot distinguish arrested division from balanced division and death.
Match the method to the question
Direct tracking events support division/death analysis. First outcomes with incomplete follow-up have their own mode. A combination matrix can be compared with Bliss and HSA references.
In the diagram: A = 80%, B = 70% live cells relative to control. Bliss expects 56%, while the observed combination is 45%; the contrast is 11 percentage points. This does not prove a mechanism or therapeutic efficacy.
Reproducible result
Plots, numeric summaries, study context and limitations stay in the report. The archive contains inputs, results, calculation code and checksums so the calculation can be repeated.
CellFate My projects is stored in this browser. Export JSON to transfer it. The report helps choose the next research question, not prescribe treatment.
What the data can support
The service does not turn ATP into proof of death, similar curves into a mechanism or lower tumor-cell counts into treatment advice. Available conclusions depend on measurements and experimental design.
Methods, sources and limitations ↗Fibroblasts, cancer-cell response, dynamics and combinations. Ready-made examples need no registration.
Choose an example ↗