begin diffbind templates

bcbio · Sep 19, 2024 · f517d42 · f517d42
1 parent 04f41f4
commit f517d42
Show file tree

Hide file tree

Showing 13 changed files with 59 additions and 0 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
diff --git a/LICENSE b/LICENSE
diff --git a/LICENSE.md b/LICENSE.md
diff --git a/NAMESPACE b/NAMESPACE
diff --git a/NEWS.md b/NEWS.md
diff --git a/README.md b/README.md
diff --git a/_pkgdown.yml b/_pkgdown.yml
diff --git a/inst/templates/chipseq/diffbind/chipseq_peakanalysis_H3K27Ac.csv b/inst/templates/chipseq/diffbind/chipseq_peakanalysis_H3K27Ac.csv
@@ -0,0 +1,25 @@
+sample,fastq_1,fastq_2,antibody,control,genotype
+WT_H3K27ac_ChIPseq_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780292_SRR6823782.fastq.gz,,H3K27ac,WT_H3K27ac_input_REP1,WT
+WT_H3K27ac_ChIPseq_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780292_SRR6823783.fastq.gz,,H3K27ac,WT_H3K27ac_input_REP1,WT
+WT_H3K27ac_ChIPseq_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780292_SRR6823784.fastq.gz,,H3K27ac,WT_H3K27ac_input_REP1,WT
+WT_H3K27ac_input_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780293_SRR6823785.fastq.gz,,,,WT
+WT_H3K27ac_input_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780293_SRR6823786.fastq.gz,,,,WT
+WT_H3K27ac_ChIPseq_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780294_SRR6823787.fastq.gz,,H3K27ac,WT_H3K27ac_input_REP2,WT
+WT_H3K27ac_ChIPseq_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780294_SRR6823788.fastq.gz,,H3K27ac,WT_H3K27ac_input_REP2,WT
+WT_H3K27ac_ChIPseq_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780294_SRR6823789.fastq.gz,,H3K27ac,WT_H3K27ac_input_REP2,WT
+WT_H3K27ac_input_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780295_SRR6823790.fastq.gz,,,,WT
+WT_H3K27ac_input_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780295_SRR6823791.fastq.gz,,,,WT
+WT_H3K27ac_ChIPseq_REP3,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780296_SRR6823792.fastq.gz,,H3K27ac,WT_H3K27ac_input_REP3,WT
+WT_H3K27ac_input_REP3,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780297_SRR6823793.fastq.gz,,,,WT
+cKO_H3K27ac_ChIPseq_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780298_SRR6823794.fastq.gz,,H3K27ac,cKO_H3K27ac_input_REP1,cKO
+cKO_H3K27ac_ChIPseq_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780298_SRR6823795.fastq.gz,,H3K27ac,cKO_H3K27ac_input_REP1,cKO
+cKO_H3K27ac_ChIPseq_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780298_SRR6823796.fastq.gz,,H3K27ac,cKO_H3K27ac_input_REP1,cKO
+cKO_H3K27ac_input_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780299_SRR6823797.fastq.gz,,,,cKO
+cKO_H3K27ac_input_REP1,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780299_SRR6823798.fastq.gz,,,,cKO
+cKO_H3K27ac_ChIPseq_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780300_SRR6823799.fastq.gz,,H3K27ac,cKO_H3K27ac_input_REP2,cKO
+cKO_H3K27ac_ChIPseq_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780300_SRR6823800.fastq.gz,,H3K27ac,cKO_H3K27ac_input_REP2,cKO
+cKO_H3K27ac_ChIPseq_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780300_SRR6823801.fastq.gz,,H3K27ac,cKO_H3K27ac_input_REP2,cKO
+cKO_H3K27ac_input_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780301_SRR6823802.fastq.gz,,,,cKO
+cKO_H3K27ac_input_REP2,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780301_SRR6823803.fastq.gz,,,,cKO
+cKO_H3K27ac_ChIPseq_REP3,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780302_SRR6823804.fastq.gz,,H3K27ac,cKO_H3K27ac_input_REP3,cKO
+cKO_H3K27ac_input_REP3,s3://hcbc-seqera/results/sra_peakanalysis_GSE111661/fastq/SRX3780303_SRR6823805.fastq.gz,,,,cKO
diff --git a/inst/templates/chipseq/diffbind/params_diffbind-example.R b/inst/templates/chipseq/diffbind/params_diffbind-example.R
@@ -0,0 +1,8 @@
+# info params
+
+
+# Example data
+coldata_fn='chipseq_peakanalysis_H3K27Ac.csv'
+peaks_dir = '~/s3_results/chipseq_peakanalysis_h3k27ac_narrow/bowtie2/mergedLibrary/macs2/narrowPeak/'
+bam_dir = '~/s3_results/chipseq_peakanalysis_h3k27ac_narrow/bowtie2/mergedLibrary/'
+peak_caller = 'macs'
diff --git a/inst/templates/chipseq/information.R b/inst/templates/chipseq/information.R
diff --git a/inst/templates/chipseq/libs/load_data.R b/inst/templates/chipseq/libs/load_data.R
diff --git a/inst/templates/chipseq/libs/make_diffbind_samplesheet.R b/inst/templates/chipseq/libs/make_diffbind_samplesheet.R
@@ -0,0 +1,26 @@
+library(tidyverse)
+library(tools)
+
+source('../libs/load_data.R')
+source('params_diffbind-example.R')
+
+coldata <- load_coldata(coldata_fn)
+
+bam_files <- data.frame(bam = list.files(bam_dir, pattern = '.bam$', full.names = T)) %>%
+  mutate(sample = sub("\\..*", "",basename(bam)))
+
+peak_files <- data.frame(Peaks = list.files(peaks_dir, pattern = 'Peak$', full.names = T)) %>%
+  mutate(SampleID = sub("\\..*", "",basename(Peaks))) %>%
+  mutate(SampleID = gsub('_peaks', '', SampleID))
+
+coldata_for_diffbind <- coldata %>% 
+  filter(!is.na(control) & control != '') %>%
+  # select(-description) %>%
+  dplyr::rename(ControlID = control, SampleID = sample) %>% 
+  separate(SampleID, into = c('sample', 'Replicate'), remove = F, sep = '_REP') %>%
+  mutate(peakCaller = peak_caller) 
+
+samplesheet <- coldata_for_diffbind %>%
+  left_join(bam_files %>% select(SampleID = sample, bamReads = bam), by = 'SampleID') %>%
+  left_join(bam_files %>% select(ControlID = sample, bamControl = bam), by = 'ControlID') %>%
+  left_join(peak_files, by = 'SampleID')
diff --git a/inst/templates/chipseq/readme.md b/inst/templates/chipseq/readme.md