10x Genomics
Chromium Single Cell CNV

Cell Ranger DNA1.0, printed on 11/21/2024

Cell Ranger DNA Algorithms Overview

The algorithms in the cellranger-dna cnv pipeline take as input

• Paired-end sequencing reads from a Chromium Single cell CNV library
• A reference genome

and produces as output

• Copy number calls for each cell
• A hierarchical clustering of cells based on mutual similarity
• Copy number calls for each group of cells defined by the hierarchical clustering

The algorithms can be organized into:

• Preprocessing
  • Strip the 10x barcode from each read pair
  • Align the reads to the reference genome
  • Identify barcodes that correspond to cells
  • Compute a coverage profile matrix for each cell at 20 kilobase (kb) genome resolution
  • Determine the mappable regions of the genome by simulation
• Single cell copy number calling
  • Estimate the copy number for each cell over the mappable regions of the genome
  • Impute copy number values over the non-mappable regions
• Clustering
  • Build a hierarchical clustering of cells using the single cell copy number calls
  • Construct a read coverage profile matrix for each group of cells defined by the hierarchical clustering
  • Determine the copy number for each group of cells defined above