Mercurial > repos > artbio > snvtocnv

annotate sequenza_to_hrdtools_input.R @ 4:e299c477b475 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit 10edea05400e21ba43942bd0cc582153d60ae2d9"
author artbio
date Sat, 20 Feb 2021 09:10:06 +0000
parents 0e54da14f831
children e66b91475343
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
1 # load packages that are provided in the conda env
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
2 options(show.error.messages = F, error = function() {
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
3 cat(geterrmessage(), file = stderr()); q("no", 1, F) })
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
4 Sys.setenv(TZ = "Pacific/Auckland") # turnaround the tidyverse bug "In OlsonNames() : no Olson database found"
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
5
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
6 library(optparse)
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
7 library(tidyverse)
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
8
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
9 option_list <- list(
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
10 make_option(
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
11 c("-i", "--input"),
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
12 default = NA,
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
13 type = "character",
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
14 help = "Path to Sequenza output segments file"
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
15 ),
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
16 make_option(
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
17 c("-o", "--output"),
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
18 default = NA,
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
19 type = "character",
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
20 help = "output file, to be used as input for HRDetect"
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
21 ),
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
22 make_option(
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
23 c("-s", "--solutions"),
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
24 default = NA,
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
25 type = "character",
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
26 help = "Path to Sequenza list of alternative solutions"
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
27 )
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
28 )
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
29
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
30 opt <- parse_args(OptionParser(option_list = option_list),
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
31 args = commandArgs(trailingOnly = TRUE))
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
32
4
e299c477b475 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit 10edea05400e21ba43942bd0cc582153d60ae2d9"
artbio
parents: 0
diff changeset
33 sequenza_data <- as_tibble(read.delim(opt$input, header = TRUE))
e299c477b475 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit 10edea05400e21ba43942bd0cc582153d60ae2d9"
artbio
parents: 0
diff changeset
34 solutions_data <- as_tibble(read.delim(opt$solutions, header = TRUE))
0
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
35
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
36
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
37 ploidy <- round(solutions_data$ploidy[1])
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
38 cellularity <- solutions_data$cellularity[1]
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
39
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
40 reformatted <- sequenza_data %>%
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
41 select(
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
42 chr = chromosome,
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
43 start = start.pos,
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
44 end = end.pos,
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
45 copynumber = CNt,
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
46 A, B
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
47 ) %>%
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
48 mutate(
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
49 ploidy = ploidy,
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
50 cellularity = cellularity,
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
51 lohtype = case_when(
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
52 copynumber == 0 ~ "HOMD",
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
53 B == 0 & A == ploidy ~ "NLOH",
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
54 B == 0 & A < ploidy & A > 0 ~ "DLOH",
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
55 copynumber > ploidy & A > B ~ "ASCNA",
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
56 copynumber > ploidy & A == B ~ "BCNA",
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
57 TRUE ~ "HET"
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
58 )
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
59 )
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
60
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
61 message("Preview of output:")
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
62 print(reformatted)
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
63
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
64 reformatted %>%
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
65 write_tsv(opt$output)
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
66
0e54da14f831 "planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/snvtocnv commit fd049d0796cbfaa6c41a7d80e84d3a734b30acc8"
artbio
parents:
diff changeset
67 message(sprintf("Output written to %s", opt$output))