v1.0.0

Final release of 'speclet'

While incomplete, this project was a strong proof-of-concept for the development of a complex Bayesian linear model fit to large-scale CRISPR screen data.

---

Below is a collection of all the PRs for this repo as a way of recognizing the work that went into the research project.

What's Changed

• Model design experimentation by @jhrcook in #1
• Update master with subset-data branch by @jhrcook in #5
• Python type hints by @jhrcook in #6
• Data subset model by @jhrcook in #11
• Data subset model (merge to fix sync issues with GitHub) by @jhrcook in #13
• Experimentation with modeling a subset of the data by @jhrcook in #14
• CRC model experimentation and workflow proof-of-concept by @jhrcook in #16
• Tests against modules in analysis by @jhrcook in #19
• Remove old cruft from attempting to fit a mimic of CERES by @jhrcook in #21
• CRC model 1 to model-fitting pipeline by @jhrcook in #26
• Use Papermill to parameterize model report notebooks by @jhrcook in #27
• Project reorganization by @jhrcook in #29
• Reorganize pipeline by @jhrcook in #30
• Design CRC Model 2 by @jhrcook in #31
• PPC LOO-CV by @jhrcook in #34
• Fix many logging issues by making a separate module. by @jhrcook in #35
• PyMC3 models as classes by @jhrcook in #38
• Update pipeline for object-oriented model fitting by @jhrcook in #39
• Create the first in a series of CERES mimicking models by @jhrcook in #40
• Simulation based calibration pipeline by @jhrcook in #41
• Flake8 code formatting by @jhrcook in #43
• Optional copy number covariate in CERES model by @jhrcook in #44
• Add an optional varying intercept for sgRNA | gene to CERES by @jhrcook in #45
• Update and run pipelines by @jhrcook in #48
• Add and fit a new model: SpecletOne by @jhrcook in #50
• Refactor model system by @jhrcook in #51
• Update pipelines by @jhrcook in #52
• Add SpecletTwo model by @jhrcook in #53
• analysis: fit SpecletOne and Two by @jhrcook in #54
• Optional KRAS covariate in SpecletTwo and reorganize model structure of fitting pipeline by @jhrcook in #55
• SpecletThree and Four by @jhrcook in #56
• Separate mcmc chains by @jhrcook in #58
• Non-centered parameterization of SpecletThree (PoC) by @jhrcook in #62
• Non-centered parameterization of SpecletFour by @jhrcook in #64
• Refactor the munging process into a single Snakemake pipeline by @jhrcook in #65
• Update data with Project SCORE and fix some known issues. by @jhrcook in #66
• Add SpecletFive and SpecletSix models by @jhrcook in #67
• Update SBC pipeline by @jhrcook in #68
• Build SpecletSeven as a deep hierarchical model by @jhrcook in #73
• Refactor SBC and model-fitting pipelines by @jhrcook in #74
• Update project documentation by @jhrcook in #77
• Improved efficiency of CI by @jhrcook in #78
• SBC refactor by @jhrcook in #79
• Snakemake pipeline build in nightly project build CI by @jhrcook in #85
• SBC limited run by @jhrcook in #86
• Test sprint (increase coverage over some modules) by @jhrcook in #87
• Model standardization refactor by @jhrcook in #88
• Configuration-driven system for entire project by @jhrcook in #92
• Munge pipeline refactor by @jhrcook in #97
• Fit models by @jhrcook in #99
• Sp7 parameterizations by @jhrcook in #104
• SBC uniformity test by @jhrcook in #105
• SpecletSimple model for easy experimentation by @jhrcook in #106
• Update environments by @jhrcook in #107
• Fix theano lock issue by @jhrcook in #108
• SpecletSeven SBC by @jhrcook in #111
• Read count data by @jhrcook in #113
• Experimentation with a Negative Binomial model by @jhrcook in #115
• Update data by @jhrcook in #122
• Mypy updates and other small chores by @jhrcook in #126
• Revisit Negative Binomial models with re-freshed data by @jhrcook in #130
• feat: standardize total read counts table col names by @jhrcook in #134
• Data manager re-write with optional dask backend by @jhrcook in #135
• style: address code review comments by @jhrcook in #136
• Speclet eight by @jhrcook in #137
• Refactor configuration system by @jhrcook in #138
• SpecletEight fitting by @jhrcook in #141
• Strict model configuration interface by @jhrcook in #142
• Build and test SpecletNine model by @jhrcook in #144
• Refactor modeling system and pipelines for using Stan by @jhrcook in #147
• Python environment update by @jhrcook in #148
• Negative binomial models by @jhrcook in #149
• Larger CRC/Bone subsample by @jhrcook in #150
• Fix Theano BLAS warning by @jhrcook in #154
• Add more pieces to the hierarchical negative binomial model by @jhrcook in #157
• Posterior predictions on new data by @jhrcook in #158
• Oncogene covariate by @jhrcook in #159
• Clean by @jhrcook in #160
• Numpyro PyMC JAX backend by @jhrcook in #162
• Refactor SLURM resources and run model-fitting pipeline by @jhrcook in #165
• Data subset with colorectal, pancreas, and esophagus lineages by @jhrcook in #166
• Attempt at modeling the colorectal, pancreas, and esophageal lineages by @jhrcook in #168
• Small chores by @jhrcook in #169
• Address CI of pytest suite by @jhrcook in #171
• Single lineage model interface by @jhrcook in #172
• Single lineage model (and some other additions) by @jhrcook in #170
• GPU support in the model fitting pipeline by @jhrcook in #179
• Simplify the model to help with sampling by @jhrcook in #176
• Model version and information by @jhrcook in #185
• Varying chromosome effects nested under cell line by @jhrcook in #187
• Update model fitting reports by @jhrcook in #188
• Expand the analysis to all lineages and s...