BART-Seq pipeline

This is a pipeline for BART-Seq implemented with Snakemake

The primer design code and an older version lives in theislab/bartSeq.

Entry point

The pipeline can be run via snakemake [-j 4] [-s …/bartseq/Snakefile] [-d …/mydata], where -j specifies the number of threads, and the other parameters default to ./Snakefile and ., respectively.

Within the data directory, the following structure is expected:

• in/

  • reads/
    • <libname>_R1_001.fastq.gz
    • <libname>_R2_001.fastq.gz
    • optionally additional libraries…
  • amplicons.fa or amplicons/<libname>.fa for all libraries
  • barcodes.fa or barcodes/<libname>.fa for all libraries

• config.yml – A file with the defaults

  amplicon-min-length: null  # You can set an integer like 70
  allow-mismatch:      True  # You can set this to False

Through the way Snakemake works, you need to create this file. leave it empty to use the defaults.

Output

The pipeline creates a process and an out directory.

The out directory contains plots and summary spreadsheets.

process/3-tagged contains the tagged FASTQ reads without alignment, but you can add a tag for the mapped amplicon by executing e.g. python -m bartseq browse NGS16 Lib1_S1_L001 | gzip >Lib1.fq.gz

Command line interface

python -m bartseq tag [<options>] [in_1] [out_1]

in_1

Read1 file to read from. Supported compression: see --in-compression

out_1

Read1 file to write to. Supported compression: see --out-compression

--in-2 IN_2	Read2 file to read from. Supported compression: see --in-compression
--out-2 OUT_2	Read2 file to write to. Supported compression: see --out-compression
--bc-file=BC_FILE, -b BC_FILE
	Barcode file in the format <ID> <Sequence> (with header)
--stats-file=STATS_FILE, -s STATS_FILE
	File to write final stats to (in JSON format)
--bc-table=BC_TABLE, -B BC_TABLE
	File name for the HTML table of barcode mismatches
--total=TOTAL, -t TOTAL
	Number of fastq records in file. “0” means no progressbar
--len-primer=LEN_PRIMER, -p LEN_PRIMER
	Primer length for stats
--len-linker=LEN_LINKER, -l LEN_LINKER
	Linker length to cut out
--in-compression=<gz|xz|bz2>, -i <gz|xz|bz2>
	Specify compression if reading from stdin or a file with unusual suffix
--out-compression=<gz|xz|bz2>, -o <gz|xz|bz2>
	Specify compression if writing to stdout or a file with unusual suffix
--dry-run, -n	Only print what would be done and exit

python -m bartseq count [<options>] data_dir [library]

data_dir

Data directory to read from. Needs to have the directories “./process/{3-tagged,4-mapped}” filled.

library

Library name. E.g. “Lib1_S1_L001” for input files named “Lib1_S1_L001_R{12}_001.fastq.gz”. Omittable if only one library exists.

--no-mismatch	Ignore barcodes with mismatches while counting.
--both	Print the count results for both to stdout. Default: Write to “./process/5-counts” instead
--one	Print the count results for one to stdout. Default: Write to “./process/5-counts” instead

python -m bartseq browse [<options>] data_dir [library] [out]

data_dir

Data directory to read from. Needs to have the directories “./process/{3-tagged,4-mapped}” filled.

library

Library name. E.g. “Lib1_S1_L001” for input files named “Lib1_S1_L001_R{12}_001.fastq.gz”. Omittable if only one library exists.

out

FASTA file to write to. Supported compression: see --out-compression

--out-compression <gz|xz|bz2>, -o <gz|xz|bz2>
	Specify compression if writing to stdout or a file with unusual suffix

Data and statistics

Read structure

• trash
• 3nt protection CCA ()
• 8nt barcode (known from set)
• Linker (one for left bcs, one for right bcs)
• Primer + Rest of Amplicon

Interesting Statistics

Make statistics: How many reads have a barcode, ...

from reads tagged with info:

• Barcode available?
• Trash before bc?
• Where bc?
• Concatamere? (bc[-bc-bc…]-linker-primer)
• Which nucleotides where bc should be?
• Amplicon maps to which gene?

Possible Problems

• No Amplicons: Only bc and linker
• Amplicon quality bad at the end
• Trash at the beginning
• Barcodes can have mismatches