This repository contains code and data from Koche et al. Extrachromosomal circular DNA drives oncogenic genome remodeling in neuroblastoma Nat Genet (2020).

Data

Circle Calls

• Data/CircleseqCircles.bed contains Circle-seq circle calls for our dataset
• Data/WGSinferredCircles_eccDNA.txt contains WGS-inferred circle calls for our dataset (eccDNA only)
• Data/WGSinferredCircles_ecDNA.txt contains WGS-inferred circle calls for our dataset (ecDNA only)

WGS Rearrangement Calls

• Data/MergedSV_StrictestFiltering_CircleSeqCircleAnnotation.txt are the merged and filtered interchromosomal rearrangement calls (see Structural Variant Merging) which were overlapped with Circle-seq circle calls for classification (available upon request)
• Data/MergedSV_StrictestFiltering_WGSCircleAnnotation.txt are the merged and filtered interchromosomal rearrangement calls (see Structural Variant Merging) which were overlapped with WGS-inferred circle calls for classification (available upon request)
• Data/PedPanCanSVs.csv contains structural variants in the DKFZ Pediatric Pan-Cancer dataset published by Gröbner et al. 2018 (data downloaded from the corresponding R2 platform)
• Data/PedPanCanMeta.csv contains metadata such as the tumor entity for the DKFZ Pediatric Pan-Cancer dataset

Expression Data

• Data/berlin_cohort_rnaseq_fpkm.txt is RNA-seq data for the Berlin neuroblastoma cohort (available upon request)
• Data/peifer_54nb_fpkms.txt is RNA-seq data published by Peifer et al.

Clinical Data

• Data/ClinicalData.csv specifies risk group and survival data

Other

• Data/BroadIntervals/b37-[...]_MinusBlckLst.txt specifies the mappable, non-blacklisted genome we use for randomization analyses

Code

Structural Variant Merging

• script_merge_translocations[...].py collapses interchromosomal structural variants from different callers
• script_filter_mapqBAM_3.0.py filters collapsed interchromosomal structural variants using common criteria
• script_merge_intraSVs_[...].py collapses intrachromosomal structural variants from different callers
• script_filter_atleast2callers_2.0.py filters collapsed intrachromosomal structural variants using common criteria
• script_merge_circularized_regions_[...].py overlaps interchromosomal structural variants with circle calls and classifies accordingly
• script_merge_circularized_regions_[...]_intraSV_[...].py overlaps intrachromosomal structural variants with circle calls and classifies accordingly
• script_search_integrations_[...].py searches for two-breakpoint patterns indicative of circle integration into a chromosome

Detection of Clusters of Rearrangements

Tree-shaped rearrangement regions are referred to as Palm Trees throughout the code.

• RunAll.R runs scripts in the correct order
• Parse*.R and Prepare*.R are several scripts to read data from different sources and create tidy representations for further analysis
• CallPalmTrees.R uses output of structural variant merging to call palm trees (a.k.a. tree-shaped rearrangements or clusters of rearrangements)
• CallPalmTreesPedPanCan.R calls palm trees on published data from Gröbner et al. 2018 (using the same settings as in CallPalmTrees.R)

Basic Rearrangement Cluster Statistics

• CircosPlots.R plots the identified rearrangements and palm tree regions
• PalmTreeStackPlot.R creates plots that integrate all CN, SV and Circle-seq information
• GeneralPalmTreeStatistics.R analyses number and length distributions of palm trees
• PalmTreeDensity.R plots genome-wide recurrence of palm trees
• PedPanCanStats.R explores palm tree occurrence in PedPanCan dataset
• CompareBerlinCohortToPedPanCan.R compares palm tree prevalence between the Peifer/Berlin dataset and the PedPanCan neuroblastoma cases
• AnalyseNBCallPalmTreesRandomization.Rmd explores 500 synthetic datasets for our cohort and the PedPanCan dataset respectively. Synthetic datasets were obtained by randomizing breakpoint positions during palm tree calling. This is used to estimate false-discovery rates due to the number of rearrangements per sample.

Genes and Gene Expression

• PalmTreeGenes.R analyzes which genes can be found within palm tree regions
• PalmTreeTargetsCloseToOncogenes.R statistically analyses whether palm tree target sites are significantly enriched in the neighborhood of cancer-related genes.
• ExpressionAnalysis.R screens for deregulated genes close to palm tree associated rearrangements

Association with circular DNA

• RegionSampling.R contains sampling methods to randomly sample regions from a masked or unmasked genome
• Overlap[...]CircleCalls.R statistically analyses the degree of overlap between palm tree regions and eccDNA / ecDNA inferred from WGS / Circle-seq
• MergeWGSCSOverlapPlots.R merges results from the preceding scripts for integrative plotting

Clinical Relevance

• ClinicalData.R explores the prognostic significance of palm trees

Breakpoint Analysis

• CircleJunctionAnalyseSvabaBreakpoints.R investigates sequence characteristics reconstructed Circle-seq circle junctions
• SVAnalyseSvabaBreakpoints.R investigates sequence characteristics interchromosomal rearrangements calls by Svaba
• MemeAnalysis.sh runs MEME on several breakpoint-associated sequences
• myHomology.R obtains microhomology estimates for accurate breakpoint coordinates from the reference genome
• Repeats.R tests for association of breakpoints with repetitive regions

Contact

If you have any questions concerning code or data, please do not hesitate to contact us at henssenlab@gmail.com.