differentially methylated regions (DMR)

This version now has a complete pipeline for running EWAS, with DMR and DMP functions.

What's Changed since last release

• Added a differentially methylated regions (DMR) functions that takes the output of the diff_meth_pos (DMP) function.
  • DMP maps differences to chromosomes; DMR maps differences to specific genomic locii, and requires more processing.
  • upgraded methylprep manifests to support both old and new genomic build mappings for all array types.
    In general, you can supply a keyword argument (genome_build='OLD') to change from the new build back to the old one.
  • Genome annotation. But won't work with mouse array.
  • DMP integrates the combined-pvalues package (https://pubmed.ncbi.nlm.nih.gov/22954632/)
  • DMP integrates with UCSC Genome (refGene) and annotates the genes near CpG regions.
  • Annotation includes column(s) showing the tissue specific expression levels of relevant genes (e.g. filter=blood)
• to_BED provides output BED for export to other genomic analysis tools
• fixed methylize.diff_meth_pos linear regression. upgraded features too
  • Support for including/excluding sex chromosomes from DMP (probe2chr map)
• dotted manhattan_plot sig line is Bonferoni corrected (pass in post_test=None to leave uncorrected)
  • probes sorted by MAPINFO (chromosome location) instead of FDR_QValue on manhattan plots now
• better test coverage in general

Full Changelog: v0.9.0...v0.9.9

Initial release with differentially methylated probe (DMP) detection and plots

v0.9 initial release (#7)

* Create requirements.txt

* Create test_analysis.py

* Create index.rst

* Create mds.py

* Update mds.py

* Update mds.py

* Create __init__.py

* Update __init__.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* Update mds.py

* integrating EWAS features from Alanna Durkin (#5)

* Add files via upload

>Notebook running bumphunter in R with variations in parameters to see how sensitive the DMRs found are to different settings like cutoff percentile values, maximum cluster size, preprocessing method, and using Beta vs M as the methylation measure in the model.

* Add files via upload

>Python notebook translating R code for bumphunter step by step before formalizing functions.

* Adding new file with python translation of minfi's dmpFinder function

* Adjusting bumphunter development branch

* Adjusting EWAS DMP finding development branch

* Added binary data input and formatting

* Added OLS regression for continuous phenotype data

* Added logistic regression

* Changed input from specifying phenotype data type to specifying regression method

* Formalized regression statistics results into output dataframe

* Fixed some bugs in logistic regression and added testing notebook

* Added handler for perfect separation errors encountered for some probes in logistic regression.

* Fully working logistic regression for pandas DF

* Working linear regression

* Added parallelization, but not parallelizing correctly yet

* Changed data structures for pool.map multiprocessing but still issues with results being overwritten

* Successful parallellization with joblib for linear regressions using pandas DF

* Successful parallellization of logistic regression with joblib

* Volcano plot function added

* Manhattan plots and refactoring stuff

* manhattan plot and viz options

* export data

* docs

* tweaked helpers

* getting imports working

* Update README.md

* docs and installer

* readthedocs

* debugging readthedocs

* Update README.md

* Create bumphunter.md

* readthedocs

* docs

* docs

* docs

* docs

* docs

* tests failed

* testing