spatial_driver_co-detection_paper

This is the repository for the spatial driver detection paper: "Spatial Co-mapping of Cancer Mutations and Gene Expression at Subcellular Resolution".

Probes were designed using the code found at Variant_site_flanking_sequence_and_probe_compatibility_scripts/Script_for_pulling_SNP_flanking_sites_v6.R

• This code runs using the packages installed in the r4.2 environment yaml file: spatial_driver_co-detection_paper/r4.2_env_used_to_generate_probe_flanks_last_update_20240201.yml
  • it relies on the following packages specifically:

  r-base
  biomaRt
  BSgenome.Hsapiens.UCSC.hg38
  GenomicFeatures
  GenomicRanges
  stringi
  

• Input targeted sequences are provided at spatial_driver_co-detection_paper/Variant_site_flanking_sequence_and_probe_compatibility_scripts/input_variant_targets.tsv
• Output targeted sequences are provided at spatial_driver_co-detection_paper/Variant_site_flanking_sequence_and_probe_compatibility_scripts/input_variant_targets_flanks.tsv
• Flank sequences require manual review and curration as outlined in the methods. that is present here: spatial_driver_co-detection_paper/Variant_site_flanking_sequence_and_probe_compatibility_scripts/manual_indel_flank_site_check.xlsx

Analysis for this paper relied upon the packages installed in the r4.3.2 environment yaml file: spatial_driver_co-detection_paper/r4.3.2_seurat5.0.1_for_analysis_20250116.yml

• Installation can be done via conda env create -n scrublet --file r4.3.2_seurat5.0.1_for_analysis_20250116.yml
  • analysis relied upon the following packages:

  Seurat_5.0.1
  SeuratObject_5.0.1
  arrow_14.0.1
  future_1.33.1
  optparse_1.7.4
  magrittr_2.0.3
  tidyverse_2.0.0
  epitools_0.5-10.1
  effectsize_0.8.9
  ggrepel_0.9.5
  ggrastr_1.0.2
  ggpubr_0.6.0
  ComplexHeatmap_2.18.0
  circlize_0.4.16
  RColorBrewer_1.1-3
  grid_4.3.2
  Matrix_1.6-5
  viridisLite_0.4.2
  patchwork_1.2.0
  scales_1.3.0
  scCustomize_3.0.0
  enrichR_3.2
  reshape2_1.4.4
  ggpmisc_0.5.5
  ggh4x_0.2.8
  

  • the libaries ComplexHeatmap and Circlize were installed directly in R with BiocManager rather than conda.
• Code related to analysis from the paper can be found in the spatial_driver_co-detection_paper/plotting_snippets/ folder. The worklogs in each folder describe how the code/tables were used and how code was run.

Other tools used in the analysis of this paper include:

Copy number calling was run with GATK4SCNA.

• This relies on two docker image: austins2/gatk4scna:v1.1 and austins2/ggplot_gatk4scna:v.2024.08.19.
• It is built for running on a HPC cluster running the IBM LSF job scheduler.
• The worklog for how it was run on the data for this publication can be found at spatial_driver_co-detection_paper/bulk_WES_cnv/GATK4_somatic_cnv_detection_worklog.sh

Scrublet for new snRNA-seq/multiome samples was run with automated_scrublet.

• the environment for scrublet automation can be found on the scrublet repository. A docker image is also provided.
• Example for the stand alone and docker image can be found there.
• The worklog for running scrublet on the new snRNA-seq data present in this cohort can be found at spatial_driver_co-detection_paper/snRNA_scRNA_preprocessing/object_generation_worklog.sh
• Some additional python scripts used the environment found on the scrublet repository. Where applicable this is noted in the worklog.

H&E images were aligned with HEX-SIFT as outlined in the worklog file.

• The envrionment and installation instructions for HEX-SIFT image alignment can be found in the above HEX-SIFT repository.
• The worklog for running HEX-SIFT on each sample can be found at spatial_driver_co-detection_paper/HE_image_alignment/HEX-SIFT_alignment_worklog.sh
• Depending on the input image file this can take up to 400GB of memory.

Mutation mapping in matching snRNA-seq data relied on scVarScan and custom scripts for post-processing as outlined in the worklog file.

• The version of perl installed on the server that scVarScan was run on was v5.26.2. The post-processing scripts run with the conda python 3.9.6 environment from automated_scrublet active.
• The worklog for running scVarScan can be found at spatial_driver_co-detection_paper/snRNA_mutation_mapping/matching_snRNA_mutation_mapping_worklog.sh
• post processing scripts are located in spatial_driver_co-detection_paper/snRNA_mutation_mapping

Variant calling (snvs/indels) was run using Somaticwrapper and it's implementation in the PECGS pipeline.

Layer calculation was run using Morph

• environment for this can be found at spatial_driver_co-detection_paper/morph_layers/. Alternatively a docker image has been made available at: austins2/spatial_driver_layer.
• scripts and relevant input tables are available at spatial_driver_co-detection_paper/morph_layer_calculations/spatial_driver_layer.yml

Data Acccess