v1.6.0

Notable updates:
• Helper functions like slice_ubr_data allow for handling of data for reference libraries with 8M sequences.
• Tracking of mednorm statistic to monitor "spikiness" of data.

Full Changelog: v1.5.0...v1.6.0

v1.5.0

Major update to handle large (8M reference sequence!) library chemical mapping:
• Use of cmuts package as replacement for RNAFramework.
• Piloting use of minimap2 as potential replacement for bowtie2.
• Creation of ubr_util.py to enable some sharing of code and also user-facing util scripts like pre-indexing for bowtie2, necessary for runs with >1M reference sequences.
• Handling of HDF5 formats in ubr_split.py, ubr_run.py, ubr_merge.py, ubr_subdidivide.py scripts and last steps in MATLAB.
• Concatenation of ubr MATLAB workspaces, again to handle large datasets.
• Adding sam_to_mutdel.py to devel to pilot mutate-and-map & MOHCA-seq analysis of big libraries.

v1.4.0

Explore scaleup to runs with 4-8M reference sequences, including:

• Allow use of precomputed bowtie-build directory
• Increased time for ubr_merge
• New ubr_check_run.py allows monitoring of error and output files from cluster runs.
• New ubr_check_stats.py prints out number of reads at each stage of an ubr_run, and is run automatically at end of the run.
• Initial exploration of cmuts pipeline and i/o with HDF5 files instead of .txt.gz files.

v1.3.2

Minor updates to explore scaleup to runs with 4-8M reference sequences, including:

• Allow use of precomputed bowtie-build directory
• Increased time for ubr_merge
• New ubr_check_run.py allows monitoring of error and output files from cluster runs.
• New ubr_check_stats.py prints out number of reads at each stage of an ubr_run, and is run automatically at end of the run.
• Bug: Need to provide flag --no_collapse or this version won't run. Fix implemented in v1.4.

v1.3.0

Numerous updates:

• Major bug fix: use uint32 in final .txt files to avoid capping counts at 255!
• Get Figs to print properly by actually focusing on figures.
• Autodetect if Read1 and Read2 sequence files have overlap and are merge-able.
• New stats summary in quick_look_ubr() includes mean-to-median ratio (uniformity measure) and mean reactivity.
• Handle Ultima datasets.
• Other small useability updates and acclerations.

v1.2.0

• Example data and output figures added
• README.md documentation filled out

v1.1.0

• Primary update: adding pseudocounts for reactivity error estimates, to avoid errors of "0.0" for positions where we see no mutations.
• Use gzip for ubr_subdivide.py when dividing UBR output by sublibraries.
• Minor updates to visualizing output of quick_look_ubr

Documentation coming next...

v1.0.0

Initial release of UBR (Ultraplex-Bowtie2-RNAFramework) pipeline.

• Complete package for running data analysis on chemical mapping mutational profiling experiments for RNA.
• Parallelization of each step as appropriate for a SLURM cluster
• Stress-tested on runs including up to 1M distinct sequences, up to 5B paired end reads.
• Most code is in Python, wrapping around other packages.
• MATLAB needed for final steps including data visualization.

Important TODOs: (1) Fill out README.md, and (2) provide test example.