In cmatKhan/brentlabRnaSeqTools: A Collection of Functions to Interact with the Brentlab RNASeq Database

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  eval = FALSE
)

library(brentlabRnaSeqTools)
library(tidyverse) # only need dplyr and readr

Get the Metadata

See the create_set vignette/article for comments on the DB_USERNAME and DB_PASSWORD lines below

metadata_df = getMetadata(database_info$kn99_host, 
                          database_info$kn99_db_name, 
                          Sys.getenv("DB_USERNAME"), 
                          Sys.getenv("DB_PASSWORD"))

Filter

There are some experiment sets which have functions included to subset the metadata, for example createNinetyMinuteInductionSet(). But, in general, expect to have to filter the metadata yourself. This is a good resource dplyr as is R for datascience

subset_metadata = metadata_df %>% 

  filter(experimentDesign=="Timecourse", genotype1=="CNAG_00000", timePoint %in% c(0, 1440))

Note in the filter above that a better method for filtering for "wildtype" strains is by the strain number. There are clinical strains in the database which also have genotype1 == "CNAG_00000".

Creating the nextflow sample sheet

I do this from my local computer, so if I want to check if the files exist, I mount the /lts directory and then change the sequence_dir_prefix below to the path to lts_sequence from my local mount point. I then change the sequence_dir_prefix to the correct one for the cluster and print that sample_sheet out for processing.

sequence_dir_prefix = "/scratch/mblab/$USER/scratch_sequence/fastq_set"

sample_sheet = createNfCorePipelineSampleSheet(
  subset_metadata,
  "fastqFileNumber",
  sequence_dir_prefix,
  check_files_flag = FALSE
)

Move the files and check that they exist on the cluster

(repeated in "running the pipeline")

This is a frustrating step, and one that I haven't re-written a function for yet in the brentlabRnaSeqTools package. Right now, this is how I do it. If you are not a computational person in the lab, please just ask one of us to move the files for you

awk -F"," '{print $2}' sample_sheet.csv > fastq_lookup.txt

mkdir scratch_sequence/fastq_set

cat fastq_lookup.txt | while read line; do rsync -aHvR $line scratch_sequence/fastq_set/; done

# and then I move the run directories back up to the level that I want them -- please just ask if this doesn't make sense. Eventually there will be a function in the brentlabRnaSeqTools. If you have a better bash method, please tell me.

Inspect and write out the sample sheet

output_path = "/path/to/sample_sheet.csv"

write_csv(sample_sheet, output_path)

Running the nextflow pipeline

See the article/vignette "Running the nextflow pipeline"

cmatKhan/brentlabRnaSeqTools documentation built on Nov. 17, 2021, 5:47 a.m.