Announcing the 26Q3 Release

Announcements
1

DepMap_Logosize

26Q3 DepMap Release Notes

NEW CRISPR AND OMICS DATA

The DepMap team is excited to announce the addition of new genome-wide CRISPR screening and Omics data, including fully integrating 150 next-generation models data from the Neiswiender, Maffa, Brenan et al. publication. The data can now be explored in the depmap portal tools including data explorer and Celligner.

image
image966×872 35.2 KB

OTHER NEW DATASETS

HCRSeq

This collection contains DNA repair profiling of 19 cell lines using the HCR-seq technology, which employs a reporter system to profile the capacity of single cells to repair a library of diverse DNA lesions while simultaneously capturing their transcriptomes using 10x single-cell RNA-seq.

Data were processed using the Cellranger pipeline, and Demuxlet was used for SNP-based demultiplexing to recover cell line identity. Cellranger outputs were post-processed using our custom HCR-seq package to derive DNA repair metrics, available on GitHub (GitHub - getzlab/HCRseq: Code for pre-processing and analyzing data generated with HCR-seq · GitHub).as well as the results of the manuscript (GitHub - getzlab/HCRseq_manuscript_2026 · GitHub). For a full description of the methodology, see Chen et al., Nature Communications, 2026.

CRISPR PIPELINE UPDATES

  • Positive ChrY gene effect centered towards 0

    • Previously, forum user ralkallas alerted us that there is an overall positive gene effect score for Y chromosome genes. This positive gene effect was caused by the distortion in copy number correction due to the prevalence of chromosome Y loss. Chronos’ copy number correction intentionally only corrects deviations from the gene mean. Since genes in the Y chromosome are so frequently lost, their mean effects are positively shifted. . To mitigate this problem, chromosome Y genes have been mean centered to 0 before performing copy number correction.

  • Improvements in library effect correction

    • Through the 26Q1 release, the library effect regularization contained a bug that prevented it from being properly applied, resulting in an undercorrection of library batch effects. The issue has been addressed in 26Q3, which led to improvements in library effect correction in CRISPRGeneEffect.csv and ScreenGeneEffect.csv. Two other bugs affecting library gene effect estimates for genes present in only a subset of libraries have also been corrected.

  • Removal of TKOv3 and Brunello only genes

    • 96 genes that are only part of TKOv3 and/or Brunello libraries and are not part of Avana, Humagne-CD, or KY libraries have been removed from the following files:

      • CRISPRGeneEffect.csv, ScreenGeneEffect.csv, CRISPRGeneEffectUncorrected.csv, ScreenGeneEffectUncorrected.csv, CRISPRGeneDependency.csv, ScreenGeneDependency.csv, ScreenNaiveGeneScore.csv, CRISPRInferredCommonEssentials.csv

    • As there are a much smaller number of TKOv3 and Brunello external screens compared to other CRISPR libraries, the library effect for TKOv3 and Brunello could not be reasonably estimated for genes that are only part of these libraries.

OMICS PIPELINE UPDATES

  • Relative CN

    • Bug fix: Fixed a bug that resulted in some genes having unexpectedly low copy numbers. This change only affects a small number of genes and cell lines.

    • Removed unmappable genes: removed genes that have low mappability in more than 80% of the gene body.

  • Mutation dataset changes

  • GIAB/hg38 BAM patches

    • Updated mutation calls for problematic regions in hg38 reference (collapsed and duplicated regions) by realigning those regions in the analysis-ready BAM to a modified reference and re-running mutect2. This provides more reliable mutation calls in 83 genes (29 protein coding), including previously observed false negative mutation calls in U2AF1. See the patch_hg38_bam workflow for the implementation.

  • Gene list updated to v5

    • New version of gene list will be made available in the portal available for pipelines

MODEL ANNOTATIONS

New subtype calls have been added to our InferredMolecularSubtype.csv matrix (also viewable in Context Explorer and other tools as an Annotation):

  • SMARCB1 LOF
  • STAG2 LOF
  • Fusions:
    • EWSR1-WT1
    • EWSR1-ATF1
    • EWSR1-CREM
    • TCF3-HLF
    • SS18-SSX (SS18::SSX1, SS18::SSX2, SS18::SSX2B, or SS18::SSX4)
  • Breast cancer subtypes:
    • ERBB2 Amplification (ERBB2 Amp) (defined as ERBB2 CN > 4, WGS + WES merged)
    • Breast ER+ (defined as ESR1 rna > 2.5 & FOXA1 rna > 6, lineages other than Breast NA)
    • Breast PR+ (defined as PGR rna > 1.5, lineages other than Breast NA)
  • PAM50 classifications:
    • PAM50 Basal
    • PAM50 Her2
    • PAM50 LumA
    • PAM50 LumB
    • PAM50 Normal
    • Breast PAM50 Unclear (defined as having the top 2 PAM50 subtype scores are within 0.1 of each other)
    • PAM50 subtype scores can be found in OmicsGlobalSignatures.csv in the following columns:
      • Breast_PAM50_Basal_score
      • Breast_PAM50_Her2_score
      • Breast_PAM50_LumA_score
      • Breast_PAM50_LumB_score
      • Breast_PAM50_Normal_score
  • IDH Mutant (IDH mut) (including IDH1: p.R132H, p.R132C, p.R132L, p.R132S, p.R132G, p.R132V, p.R132Q, p.Y139D, p.R100A, p.R100Q, p.G97D and IDH2: p.R140Q, p.R172K, p.R172S, p.R172T, p.R172W, p.R140W, p.R172G, pR172M)
  • Amplifications:
    • MYCN (defined by relative CN > 10, WGS + WES merged)

OTHER PORTAL UPDATES

  • Generating small multiples by faceting plots
    • Users now have the ability to facet plots by any group of features, generating a set of small-multiples. Faceting is enabled for of both sets of models and sets of features (for example: genes, transcripts)
    • To generate this type of plot, enter Data Explorer and set sliders to ‘multiple’ and ‘facet’. From there, select your context to facet by.

image

  • Due to plotting constraints, we set a limit of small multiples per datatype. For example, faceting by models allows you to plot 3 and genes allows for 8 small multiple plots.

  • After generating your plot, further refinement is possible by clicking on “Choose…” in the legend and filtering by specific characteristics of the feature you are faceting by.

  • Context manager update

    • All the properties of the models being used in your analysis are now available for you to use in Data Explorer.
    • A clear use case for this update will be to load a CRISPR analysis in Data Explorer. Select your points as CRISPR screens and you’ll be able to Filter (make a context) by properties at the Model and Model Conditions levels.
    • This will allow you to filter by Organoid, Spheroids, and more.
    • Keep scrolling to see the range of metadata columns available for you to use.
DepMap Quarterly Release Notes