All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog.
- No longer calling transform() on parameters for single condition fits. See #160.
- Added init_beta_variance parameter to the Model instantiation to allow the user to initialize beta parameters by sampling a normal distribution. See #161.
- This release re-implements the joint model as a using a generalized lasso, and bit-flipping, as described in #156. Please see the issue for more detailed description about how, and why these changes were made. Note that this changes the parameters that one may get from the model including a set of beta's for each experimental condition.
- It also cleans up various TODO's in the code as checked-off in #153 <#153>.
- Fixes a bug, where the phenotype predictions for single mutants did not correctly include the bundle effects.
- Fixes and cleans various plotting bugs.
- new simulation validation analysis and plotting functions (at the time of re-submission)
- fixes bug described in `#130 https://github.com/matsengrp/multidms/issues/130`_, having to do with pandas groupby.apply 2.2.0 behavior change.
- updates python version requirements to 3.9 or newer, as 3.8 did not work with the new pandas version, 2.2.0 bug patch described above.
- supresses the cpu warning from jax.
- adds ModelCollection.add_validation_loss, ModelCollection.get_conditional_loss_df, Model.conditional_loss, and Model.get_df_loss methods, which can all be used quite easily to perform cross validation analysis.
- simply updates the ruff linting to version 0.0.289
- fixes bug #128
- fixes bug #126.
- Adds the initial working simulation notebook.
- Adds initial multidms.model_collection module with
multidms.fit_modelsfor the ability to fit multiple models across a range of parameter spaces in parallel using multiprocessing. This is inspired by the polyclonal.fit_models function. - Adds the
ModelCollectionclass for split-apply-combine interface to the mutational dataframes for a collection of models - Adds four altair plotting methods to
ModelCollection:mut_param_heatmap,shift_sparsity,mut_param_dataset_correlation, andmut_param_traceplot. - removes
utilsmodule. - Cleans up #114
- optionally removes "wts", "sites", "muts" from the mutations dataframe returned by
Model.get_mutations_df. - Changes the naming of columns produced by
Model.get_mutations_df(), in particular, it moves the condition name for predicted func score to be a suffix (as with shift, and time_seen) rather than a prefix. e.g. "delta_predicted_func_score" -> "predicted_func_score_delta".
- Fixed a bug
- caused by non-unique indices in input variant functional score dataframes.
- Made lineplot_and_heatmap() more private to remove from docs.
- Fixed bug pointed out by @jbloom #110
Model.get_mutations_df()now sets the mutation as the index- added some testing utils
- Closed a docs testing issue, thanks, @WSDeWitt !
- Cleaned Actions, again thanks to @WSDeWitt
- Fixed a bug in wildtype predictions
- Implemented QC on invalid bundle muts as pointed out by @Haddox.
- a few other minor cleanup tasks.
- First Release on PYPI