CelegansATACseq

The code for Daugherty, et al 2017 - Chromatin accessibility dynamics reveal novel functional enhancers in C. elegans

Figure|Panel|Comments|Program name ---|---|---|--- 1|A/B|General pipeline|ATACPipeline_combined.sh ||||pyadapter_trim.py ||||random_split_fastq.pl |||get consensus peaks|reRunningWithoutL1s.sh ||||splitMACS2SubPeaks.pl ||||smart_merge.py ||C|Differential, as well as 1B|30JanDiffBind_allButL1_metaPeaks_withGDNA.R ||D & E|Gene plots|examples.R |||support for the gene plots|gvizSupportFunctions_noLog.R ||F & G|How genes were connected to peaks; genes were then sorted using command line sort, and gene lists copied over to Gorilla|connectingPeaksAndTss_withConcentrations.sh |||Many of these combinations were ad hoc, so this is harder to understand|selectTerms_ggPlot2_EEvL3_allConnectedATACPeakTotalChange_GOProcess_splitup_selectGroupedTerms.R |||| S1|A|Plots from Picard in the General pipeline|From above (1 A/B) ||B||jaccardOfRegions.R ||C/D|Volcano plots|plotDifferences.24Mar2015_volcano.R ||E&F|Many of these combinations were ad hoc, so this is harder to understand|selectTerms_ggPlot2_L3vYA_allConnectedATACPeakTotalChange_GOProcess_splitup_selectGroupedTerms |||| S2|A/B|description and how each of the remaining scripts was called; mini pipeline|whatWasDone |||called in the above|trim_galore_SE.sh |||called in the above|mapqFilterAndStrandSpecificSplit.sh |||called in the above|GROseq_alignment_bowtie2.sh |||called in the above|getConsensusPeaks_2BioReps_version4.sh |||called in the above|deDup.sh |||Plotting of the ATAC and GRO-seq data|plottingPromAccessibilityVsGroSignalRefChen.R ||C/D|All the scripts for RNA-seq analysis|runningUcscAlignments.sh |||All the scripts for RNA-seq analysis|topHat2_RNAseq_alignmentForModEncodeData_36bp_PE.sh |||All the scripts for RNA-seq analysis|topHat2_RNAseq_alignmentForModEncodeData_76bpReads_SE.sh |||All the scripts for RNA-seq analysis|quantifyTophat2ResultsWithHtSeq_RefSeq.sh |||Plotting of the ATAC and RNA-seq data|See plottingPromAccessibilityVsGroSignalRefChen.R |||| 2|A|Step by step of what was run|ExactlyWhatIdid.txt |||all of the various scripts|chromHMM_scripts |||Generating the null distribution|getEnrichments_justBash.sh |||Measuring intersection with the null|createShuffledBedDirectory_mappableSubtractBlacklistAndGDNAPeaks_parallel.sh |||actually generating the bar plots|plottingAllTogetherInTheirStates.R ||B|Standard NGS plot code|Example command: ngs.plot.r -G ce10 -R bed -C ConfigVsInput.txt -O L3_HM_SignalVsInput_atL3ATACPeaks_localScale -P 0 -IN 1 -GO max -N 0.33 -VLN 0 -CO blue:red |||How the Chen TSSs were used to replace the canonical TSS|replaceFeaturesWherePossible_byName.py ||C/D|Getting the data|generatingChangeMatrixData.sh |||plotting|chreatingChromHMMStateChangeHeatMaps.R ||E|Example plot|examples.R |||| S3|A-C|See ChromHMM code above|N/A ||||plottingStackedBarPlots.R ||D/E|Getting the data|splitAndCombineHiHMMStates.sh ||||compare_hiHMMToMyChromHMM.sh ||||runningCompares.sh |||plotting|jaccardOfChromatinStatePreds.R ||F|Standard NGS plot code|See above ||G/H|See 2C/D|N/A |||| S4|A|See 2A ||B|Generating Null distribution|generatingAllDistalNegsOnScg3_withoutProms.sh ||||plottingConservationScoresWithNulls.R ||C||plottingConservationScoresVsHistoneMods_noNulls_medians.R ||D|See examples.R|for general approach |||| 3/S5|all plots| |enhsCands.R |||| 4|A|finding known motifs|findMotifsGenome_sizeGiven_narrowPeak_noL1sMetaPeakBackground_novelMotifsLO9.sh ||B|Labeling peaks|makeEEL3Labels.sh |||Getting the counts of each motif in the peaks|getMotifCounts_EEvL3DeNovoToo.sh |||Final prep of the input data|condenseCountsPerPeaks_withPreDoneNames.py |||building then predicting from the model|predictingAccessibilityChangesWithHughesAndDeNovoMotifs_gbmMetricBalAcc_LO9Peaks_loopingOverInteractionDepth_noTSSDist.R ||C|This isn't the exact code used, but this is a functio used in all predictors, so it's identical to what was then modified to make prettier for plotting|plotting_rel_importance.R ||D|Getting the data|getInsertSizesAndCountsInTfPeaksATAC100bpSummits_forAllStages.sh |||Processing the data|processingAtacInsertSizesAndEnrichmentInTfPeak100bpSummits.R |||supporting the above|ProcessingAtacInsertSizesAndEnrichmentInTfPeaks_supportFunctions.r |||Plotting|plottingATACSignalAsFunctionOfATACFragmentSize.R ||E|just the plotting|plotting_eor1_insertSize_hist.R ||F|wrapper script for DANPOS|runningDanpos.sh |||called above|wigToBedGraph.py |||Calculate the differences|nucDiffSignalInTF100bpSummits.sh |||called above|getH3NucleosomeEEL3DiffCoverageInRegions.sh |||plotting|plottingInferredNucEEL3DiffSignalInTF100bpSummits.R |||called above|plottingAtacInsertSizesAndEnrichmentInTfPeaks_supportFunctions.R |||| S6|A|See 4A for example| ||B|See 4A for example| ||C|de novo motifs were called using standard homer settings. See methods for details| ||D|plotting accuracy of model|plottingBalancedAccuracy.R |||| S7|A|See S6C for example| ||B|See 2A for example of approach| ||C/D|Plotting boxplots for each factor|plottingATACSignalAndFragmentSizeInTFChIPPeaks_v2.R ||E|See 4E| ||F|Generating the data|bedtoolsFisherTestOneVsManyWrapper.sh |||Plotting|plot_fishers_test_ORs_L3inL3.R |||| Sup Table 3||Mostly copy and paste of how I pieced together the annotation files|annotating_consensus_atac_peaks.sh |||used in the above to rename some things|combiner.py