starfish_all: starfish_all

In yanglab-computationalgenomics/Starfish: Detecting connected CGR regions and inferring mechanisms from structural variation and copy number data

starfish_all

Description

This function loads SV, CNV and gender files, identifies "connected" complex genome rearrangement (CGR) regions and complex SVs, then constructs a CGR event vs. feature matrix and infers CGR signatures based on clustering and classification.

Usage

starfish_all(
  sv_file,
  cnv_file,
  gender_file,
  prefix = "",
  genome_v = "hg19",
  cnv_factor = "auto",
  arm_del_rm = TRUE,
  plot = TRUE,
  cmethod = "class"
)

Arguments

sv_file	a SV dataframe with 8 columns: "chrom1","end1", "chrom2","end2","svtype" (DEL,DUP,h2hINV,t2tINV,TRA),"strand1" (+/-) and "strand2" (+/-),"sample". Other svtypes like INV, INS, BND are not accepted
cnv_file	a CNV dataframe with 5 columns: "chromosome","start","end","total_cn", and "sample". "total_cn" should contain absolute copy numbers.
gender_file	a sample table with 2 columns: "sample" and "gender". Gender could be "Female, "female","F","f","Male","male","M", or "m". If the gender is unknown, any other characters could be given, such as "unknown", and the gender will be inferred by the CN baseline of chromosome X
prefix	the prefix for all intermediate files, default is none
genome_v	which genome assembly was used to call SV and CNV. It should be "hg19" or "hg38", default is "hg19"
cnv_factor	the CN fluctuation beyond or below baseline to identify loss and gain fragments for samples with decimal CN, default is "auto", or users can provide a value between 0 and 1
arm_del_rm	the logical value of removing arm level deletion or not, default is TRUE
plot	the logical value of plotting "connected" CGRs, default is TRUE
cmethod	"class" based on a pre-constructed classifier from PCAWG dataset or "cluster" based on de-novo unsupervised clustering, default is "class"

Value

a list of files: $complex_sv contains complex SVs, $starfish_call contains connected CGRs, $feature_matrix contains CGR feature matrix, and $chrss_signature shows the signatures computed and inferred from PCAWG dataset. The signature classification table and plot will be generated if "class" is selected, otherwise clustering matrices and plots will be stored under "CGR_cluster" folder with K from 2 to 10.

yanglab-computationalgenomics/Starfish documentation built on July 27, 2022, 10:26 a.m.