v1.0

BiorXiv Release

v0.1.7

Full Changelog: v0.1.6...v0.1.7

v0.1.6

Fixes issue #150 and handles filtering of regions for contribution bigwig building correctly. The user no longer needs to pre-filter regions as mentioned in the wiki for generating contribution scores.

Full Changelog: v0.1.5...v0.1.6

v0.1.5

Full Changelog: v0.1.4...v0.1.5

v0.1.4

• (MAJOR) Bug in chrombpnet modisco_motifs command. seqlets was limited to 50000. If users wanted to change it to 1 million this did not happen.
• Filter peaks at edges for pred_bw command and bias pipleline. So bias evaluation now done on these filtered peaks.
• Preprocessing deafulted to use unix sort. Provided option to switch to bedtools sort.
• Provided option to use filter chromosomes option in preprocessing.

Full Changelog: v0.1.3...v0.1.4

v0.1.3

What's Changed

• fix: correct misspelling resonse->response in output files by @riyavsinha in #107
• Migrate from tensorflow-gpu to tensorflow by @austintwang in #105
• Temporary patch to fix deepish incompatibility with numpy (will replace deepdish in future versions)
• Added ISMB talk to documentation

Full Changelog: v0.1.2...v0.1.3

v0.1.2

What's Changed

• Consolidate docker layers by @annashch-insitro in #85
• fix: modify preprocessing to filter reads from chroms not in ref fasta by @riyavsinha in #87
• Added scripts to compress deepshaps, float16 on shaps and int8 on the raw
• modisco-lite version upgrade

New Contributors

• @annashch-insitro made their first contribution in #85
• @riyavsinha made their first contribution in #87

Full Changelog: v0.1.1...v0.1.2

v0.1.1

What's Changed

• Made marginal foot-printing check stringent by checking for < 0.003 across all motifs
• Fixed Chrombpnet pipeline issue in generating pdf/html files for DNase
• Moved to using bedtools sort instead of unix sort
• Set counts loss weight to 1 for both chrombpnet and bias model when reads are low - should now accomodate low read data as well
• fixed typo bug in chrombpnet modisco run
• Testing low read run by @panushri25 in #84

Full Changelog: v0.1.0...v0.1.1

pre-release-v1.2

What's Changed

WARNING: If upgrading from v1.0 or v1.1 to v1.2

Note that chrombpnet has undergone linting to generate a modular structure for release on pypi.Hard-coded script paths are no longer necessary. Please refer to the updated README (below) to ensure your script calls are compatible with v1.2

• gc-matching and preprocess by @panushri25 in #1
• fixed consistency in gc_matched_negatives by @panushri25 in #3
• Preprocessing by @panushri25 in #2
• Variant effect prediction by @panushri25 in #4
• changed output_dir to output_prefix to align with script by @kmualim in #5
• Fixed bug with merged bam file directory in data download step of tut… by @amanpatel101 in #7
• fix typos in documentation by @kmualim in #8
• Fixes to docker image and tutorial by @amanpatel101 in #9
• seed set properly to ensure reproducibility by @panushri25 in #19
• adding pseudocount to prevent div by 0 error by @annashcherbina in #13
• added commands to stop any shell script on error and return fail code by @annashcherbina in #12
• Fixed misspelling of "median_norm_jsd" by @amanpatel101 in #21
• Fixed pandas errors - added changes this time lol by @amanpatel101 in #15
• invivo footprinting changes to fit to the new format by @panushri25 in #22
• Changing marginal footprinting operate on the test chromosomes, not a single hardcoded one by @amanpatel101 in #20
• fix typo by @ziwei-75 in #23
• Fast genomewide GC binning by @austintwang in #11
• Remove "import keras" from predict_to_bigwig.py by @amanpatel101 in #24
• Updated version of Modisco in requirements.txt by @amanpatel101 in #25
• removing unused/empty chrombpnet_with_bias_bigwig.py file by @annashcherbina in #35
• added model arg to fetch_snp_predictions function by @annashcherbina in #34
• typo fix by @annashcherbina in #32
• typo fix by @annashcherbina in #33
• changed --output_dir to --output_png as the code expects a path to an… by @annashcherbina in #36
• option to provide ylim for marging footprint plotting by @annashcherbina in #27
• Faster genomewide gc counting by @suragnair in #30
• Maintain consistency with get_genomewide_gc_bins.py by @panushri25 in #38
• make num_shuffles a param and replace hard coded values of 20 by @annashcherbina in #39
• No jitter on negatives for staging by @annashcherbina in #53
• Improving modisco tomtom html documentation for staging by @annashcherbina in #52
• Staging by @annashcherbina in #54
• generate installable package for chrombpnet that corresponds to pypi release 0.1 by @annashcherbina in #26
• No jitter on negatives by @panushri25 in #37
• Improving modisco tomtom html documentation by @panushri25 in #40

New Contributors

• @panushri25 made their first contribution in #1
• @kmualim made their first contribution in #5
• @amanpatel101 made their first contribution in #7
• @ziwei-75 made their first contribution in #23
• @suragnair made their first contribution in #30

Full Changelog: https://github.com/kundajelab/chrombpnet/commits/v1.2

pre-release-v1.1

What's Changed

• gc-matching and preprocess by @panushri25 in #1
• fixed consistency in gc_matched_negatives by @panushri25 in #3
• Preprocessing by @panushri25 in #2
• Variant effect prediction by @panushri25 in #4
• changed output_dir to output_prefix to align with script by @kmualim in #5
• Fixed bug with merged bam file directory in data download step of tut… by @amanpatel101 in #7
• fix typos in documentation by @kmualim in #8
• Fixes to docker image and tutorial by @amanpatel101 in #9
• seed set properly to ensure reproducibility by @panushri25 in #19
• adding pseudocount to prevent div by 0 error by @annashcherbina in #13
• added commands to stop any shell script on error and return fail code by @annashcherbina in #12
• Fixed misspelling of "median_norm_jsd" by @amanpatel101 in #21
• Fixed pandas errors - added changes this time lol by @amanpatel101 in #15
• invivo footprinting changes to fit to the new format by @panushri25 in #22
• Changing marginal footprinting operate on the test chromosomes, not a single hardcoded one by @amanpatel101 in #20
• fix typo by @ziwei-75 in #23
• Fast genomewide GC binning by @austintwang in #11
• Remove "import keras" from predict_to_bigwig.py by @amanpatel101 in #24
• Updated version of Modisco in requirements.txt by @amanpatel101 in #25
• removing unused/empty chrombpnet_with_bias_bigwig.py file by @annashcherbina in #35
• added model arg to fetch_snp_predictions function by @annashcherbina in #34
• typo fix by @annashcherbina in #32
• typo fix by @annashcherbina in #33
• changed --output_dir to --output_png as the code expects a path to an… by @annashcherbina in #36
• option to provide ylim for marging footprint plotting by @annashcherbina in #27
• Faster genomewide gc counting by @suragnair in #30
• Maintain consistency with get_genomewide_gc_bins.py by @panushri25 in #38
• make num_shuffles a param and replace hard coded values of 20 by @annashcherbina in #39
• Improving modisco tomtom html documentation by @annashcherbina in #56

New Contributors

• @panushri25 made their first contribution in #1
• @kmualim made their first contribution in #5
• @amanpatel101 made their first contribution in #7
• @ziwei-75 made their first contribution in #23
• @suragnair made their first contribution in #30

Full Changelog: https://github.com/kundajelab/chrombpnet/commits/v1.1