Develop

[joachimwolff]

• Flake8 passes (flake8 . --exclude=.venv,.build,planemo_test_env,build --ignore=E501,F403,E402,F999,F405,E712)
• Local tests pass (py.test hicexplorer --doctest-modules)

• A new TAD prediction method: hicTADClassifier and a method to train own classifiers hicTrainTADClassifier. Thanks @AlbertLidel for the design and implementation.
• New file formats for capture Hi-C modules. chicViewpoint, chicSignificantInteractions, chicAggregateStatistic and chicDifferentialTest use now HDF5 based files. A new export script chicExportData is provided to retrieve text files from these HDF5 files.
• Implementing a few feature requests:
  • hicPlotMatrix: TADs can be visualized with hicPlotMatrix
  • hicAdjustMatrix is able to remove intra- or inter-chromosomal contacts (Plot whole genome excluding intra-chromosomal (cis) contacts (i.e., show only trans contacts)? #664, How to remove inter-chromosomal interactions from matrix #704)
  • hicValidateLocations: An option to validate TADs and to use additional data stored in cool matrices
  • hicPCA: Adding a function to select which eigenvector should be used for the output (Option to generate just one eigenvector and, separately, select which one you want #669)
  • hicConvertFormat: Adding the function to export hicpro file format and to import 2D-text files.
  • hicFindRestSites: Support of multiple restriction cut sequences (Multiple enzyme hicFindRestSites support? #659)
  • hicPlotMatrix: Option for loop locations spanning more than one bin to define if the start, center or end should be used for plotting (hicPlotMatrix --loop with regions spanning more than one bin. #640)
  • hicInterIntraTAD: A new script to compute the ratio of inter and intra TAD. ($404)
  • hicAggregateContacts: Option to consider the strand orientation (suggestion for hicAggregate and hicAverage #633)
  • hicAverageRegions: Option to consider the strand orientation (suggestion for hicAggregate and hicAverage #633)
  • hicCompareMatrices: An option to not normalize the matrices before the computation. (hicCompareMatrices without normalization #677, No norm in hicCompareMatrices #645) Thanks @lldelisle
  • hicDifferentialTAD: Adding rank sum statistics to the output (Add W rank sum statistic to hicDifferentialTAD outpu #728, Output Ranksum statistics #727). Thanks @dawe
  • hicPlotDistVsCounts: Adding a function to plot the counts vs distance in TAD regions. (Plot dist vs counts #696) Thanks @lldelisle
• Bug fixes:
  • hicCorrectMatrix: A bug that lead to wrong correction factors for the KR correction for cool files (Normalization and Correction after hic2cool #724)
  • hicDifferentialTAD: Solved multicore issue related to skipping data at the start and end of chromosomes (hicDifferentialTAD skips regions #725, ERROR:hicexplorer.hicDifferentialTAD:local variable 'right_boundary_index_target' referenced before assignmentTraceback #685)
  • hicHyperoptDetectLoops: Added an option to set if the chr prefix should be added or removed (hicHyperoptDetectLoops adds chromosome prefix by default #723)
  • hicPCA: Solving an issue if the region defined by the gene track is larger the region stored in the interaction matrix (hicPCA AttributeError : 'list' object has no attribute 'real' #655, --extraTrack only using the longest transcripts #710, hicPCA error when using gene track #716, hicPCA using Gene Track  #719)
  • hicPCA: Fixing a bug where the masking of bins was automatically applied which lead to differing matrix dimensions for the e.g. the Pearson correlation matrices (hicPCA and NaN bins #618)
  • hicBuildMatrix: Solving a bug if multiple restriction cut sites have the same dangling ends (Error in hicBuildMatrix when multiple restriction enzymes have the same dangling sequence #720)
  • hicBuildMatrix: Solving a bug that the parameter --removeSelfLigations was always set to true. Changed parameter name to --keepSelfLigations to keep the functionality. If the parameter is not set, the self ligations are removed.
  • hicBuildMatrix: If a region is specified, only the restrictionCutSite file information for that region is loaded to save memory (hicBuildMatrix - redundant memory usage #646)
  • hicConvertFormat: Fixing a bug to copy the genome annotation information in the case of a cool to cool file conversion (metadata missing after .cool to .cool conversion #657)
  • hicCorrelate: Correcting the range of colors for the heatmap (hicCorrelate does not plot the right colormap in the correlation heatmap #585)
  • hicCompartmentalization: Fixed index bug (hicComparments question #635, Fix #635 #637) Thanks @LeilyR
• Updating hicBuildMatrix to be able to work with biopython versions > 1.77. Thanks @lldelisle (new biopython version incompatible with HiCExplorer #691)