- Make sure relative fit function is defined for all model types
- Fix documentation generation on readthedocs
- Drug combination heatmaps for DIP rate (
thunor.plots.plot_drug_combination_heatmap) - Allow comparison of DIP rate and viability parameters (e.g. IC50 from each) on same plot
- Add converters (
thunor.converters) for Teicher Small Cell Lung Cancer dataset and Genomics of Drug Sensitivity in Cancer dataset - Upgrade to plotly 3.6.1
- Exclude truncated parameter values (e.g. IC50) from linear fit on scatter plots
- Fix where replicate points appear on relative DIP rate dose response curves
- Fix reading HDF5 files from a buffer, rather than a file
- Support for DIP rate calculation with only two time points
- Use control data points to calculate relative DIP rate on dose-response "curves" where no fit was available
- Document CTRP converter
- Hill coefficient plots now use log scale
-
Add rejection test for DIP curve fits based on E0:
If number of controls >=5, then reject fit if E0 > (mean(controls) + stddev(controls)) Else reject fit if E0 > 1.2 * mean(controls)
- IncuCyte Zoom format parser
- Colour by dataset on two-dataset plots
- Mann-Whitney U test on bar plots with two groups
- One-way ANOVA on box plots
- Support for 1536 well plates
- Various file parser improvements
- Sort box plots ascending order, to match bar plots
- Only use control wells from same plate(s) as experimental data wells
- Viability curves now use 3 parameter model, with upper plateau fixed at 1
- Fix axis label on viability plot
- Viability calculation and plots (end point or other single time)
- Output to CSV format
- Curve fits now identify the failure mode (whether numerical, QC failure, or statistical rejection)
- Updated documentation and tutorial
- Thunor version saved in HDF5 files to aid reproducibility
- More unit tests
- Fix DRC plot on dataset without controls
- Packaged renamed to Thunor
- Box plot of control DIP rates by plate (for QC)
- Plate map layout with DIP rate heat map
- Format drug combinations properly in plot title
- Set drug effect E_N to N/A where EC_N is N/A, e.g. for EC100
- Drug combination support in HDF5 export
- Original plate well number is now saved in HDF5 files
- Fix HDF5 export for datasets with control wells
- dataset cell_lines property now returns control-only cell lines as well as cell lines with experimental data
- Demonstrate dataset filtering options and examples in tutorial
- Fix identification of controls in drug combination CSV uploads
- Fix DIP rate overlay on timecourses when no control data are present
- User defined min/max dose for activity area/area under curve plots
- plotly_to_dataframe function for converting plotly data to pandas (and ultimately to CSV)
- Box plots for aggregating parameter values over cell lines and/or drugs
- Drug combination time course plots
- Plots comparing a parameter across two datasets
- Plots with custom IC, EC, E values
- Jupyter notebook tutorial added
- Started unit test suite
- Example data added
- fit_dip_params now returns a pandas dataframe
- EC50s out of measured range are highlighted on parameter plots
- Multiple plot layout and annotation improvements
- Control wells are now defined as any well where the dose for every drug is zero
- PyDRC data structures now support drug combinations (although this is not fully supported in the analysis functions and plot interface yet)
- Activity area as new DIP curve metric
read_hdfcan now read from buffers as well as files- DIP rate fit now returns first timepoint used in fit and y-intercept
- DIP rate overlay option for timecoure plots
- Initial version