Skip to content

Latest commit

 

History

History
256 lines (220 loc) · 9.04 KB

Demultiplexing.md

File metadata and controls

256 lines (220 loc) · 9.04 KB
Raw

Demultiplexing with TagDust 2

Multiplexing is done by adding artificial sequences to molecules, to encode their sample of origin, so that multiple samples can be sequenced together (to save costs). Demultiplexing is to retreive this information from the sequence reads, and recode it as an addition to the read name or by grouping the reads in separate files. Some multiplexing strategies are supported natively by some sequencers, but a lot of custom designs also exist.

In this example, we will install TagDust 2, download nanoCAGE data from DDBJ, and demultiplex it. These commands have been tested on a virtual machine running Linux.

If $HOME/bin is not available, create it.

mkdir -p $HOME/bin

Log out, and log in again. On Linux systems such as Debian, $HOME/bin is now in the $PATH.

If TagDust 2 is not available, install it.

Here we are downloading version 2.33. This is the first version where the option -show_finger_seq is available.

wget http://downloads.sourceforge.net/project/tagdust/tagdust-2.33.tar.gz
tar xvfz tagdust-2.33.tar.gz 
cd tagdust-2.33
./configure
make
cp src/tagdust $HOME/bin
cd

Download the sequences from DDBJ.

These are Read 1 and Read 2 from a HiSeq lane where nanoCAGE libraries have been multiplexed using a barcode that spans the first 6 bases of Read1. Illumina's TruSeq indexing system has not been used here, which is why we need to demultiplex by hand. The reason for using barcodes in Read 1 instead of TruSeq indexes is to allow for multiplexing the library construction at an early stage.

wget ftp://ftp.ddbj.nig.ac.jp/ddbj_database/dra/fastq/DRA004/DRA004180/DRX044600/DRR049557_1.fastq.bz2
wget ftp://ftp.ddbj.nig.ac.jp/ddbj_database/dra/fastq/DRA004/DRA004180/DRX044600/DRR049557_2.fastq.bz2

Demultiplex.

The sequence of Read 1 should match with BBBBBBNNNNNNNNTATAGGG, where B is a barcode base, N is a UMI base, and TATAGGG is a linker sequence that is the same in every read. The rest of the bases are cDNA sequences to be aligned on the genome.

Create an architecture file for TagDust.

The architecture files declare the structure of the reads in a machine-readable format, where B, F, S and R bloks declare barcodes, fingerprints (UMIs), spacers (linkers), and the rna sequence to be aligned. More details can be found in the user manual (the doc/User-Manual.pdf file in TagDust's source archive) or in Lassmann T., 2015.

cat > DRR049557.arch <<__END__
tagdust -1 B:ACATGA,ATCATA,CACGTG,CGATGA,GAGATA,GCTCTC,GTATGA,TCGATA,AGTAGC,ATCGCA,CACTCT,CTGAGC,GAGCGT,GCTGCA,TATAGC,CACGAT,CTGACG -2 F:NNNNNNNN -3 S:TATAGGG -4 R:N
tagdust -1 R:N
__END__

Run TagDust in paired-end mode with the architecture file.

tagdust -t 4 -show_finger_seq -ref tagdust.fa -arch DRR049557.arch -o DRR049557 DRR049557_1.fastq.bz2 DRR049557_2.fastq.bz2

-t 4 allocates 4 cores to the task. You can adapt this number to the machine where you run that command. Note that even when using multiple cores, TagDust may take hours before finishing. Look at the timestamps in the example output below for an example.

The file tagdust.fa contains sequences of nanoCAGE linkers and of Nextera sequencing linkers.

As per Illumina's Illumina Adapter Sequences Document , here is the copyright notice for the sequences that have a name starting with "Nextera".

Oligonucleotide sequences © 2016 Illumina, Inc. All rights reserved. Derivative works created by Illumina customers are authorized for use with Illumina instruments and products only. All other uses are strictly prohibited.

Rename the samples according to their barcodes.

Here is the list associating barcodes to sample names. It was constructed using the information submitted to DDBJ (https://trace.ddbj.nig.ac.jp/DRASearch/experiment?acc=DRX044600).

cat > DRR049557.samples.txt <<__END__
ACATGA HaCaT_r1
ATCATA HaCaT_r2
CACGTG HaCaT_r3
CGATGA c33a_r1
GAGATA c33a_r2
GCTCTC c33a_r3
GTATGA Hela_r1
TCGATA Hela_r2
AGTAGC Hela_r3
ATCGCA Caski_r1
CACTCT Caski_r2
CTGAGC Caski_r3
GAGCGT SiHa_r1
GCTGCA SiHa_r2
TATAGC SiHa_r3
CACGAT W12E_r1
CTGACG W12E_r2
__END__

Read the list and rename with a loop.

cat DRR049557.samples.txt |
  while read from to
  do
    prename -f "s/$from/$to/" *$from*
  done

Result

The following files should be produced:

DRR049557_BC_Caski_r1_READ1.fq
DRR049557_BC_Caski_r1_READ2.fq
DRR049557_BC_Caski_r2_READ1.fq
DRR049557_BC_Caski_r2_READ2.fq
DRR049557_BC_Caski_r3_READ1.fq
DRR049557_BC_Caski_r3_READ2.fq
DRR049557_BC_HaCaT_r1_READ1.fq
DRR049557_BC_HaCaT_r1_READ2.fq
DRR049557_BC_HaCaT_r2_READ1.fq
DRR049557_BC_HaCaT_r2_READ2.fq
DRR049557_BC_HaCaT_r3_READ1.fq
DRR049557_BC_HaCaT_r3_READ2.fq
DRR049557_BC_Hela_r1_READ1.fq
DRR049557_BC_Hela_r1_READ2.fq
DRR049557_BC_Hela_r2_READ1.fq
DRR049557_BC_Hela_r2_READ2.fq
DRR049557_BC_Hela_r3_READ1.fq
DRR049557_BC_Hela_r3_READ2.fq
DRR049557_BC_SiHa_r1_READ1.fq
DRR049557_BC_SiHa_r1_READ2.fq
DRR049557_BC_SiHa_r2_READ1.fq
DRR049557_BC_SiHa_r2_READ2.fq
DRR049557_BC_SiHa_r3_READ1.fq
DRR049557_BC_SiHa_r3_READ2.fq
DRR049557_BC_W12E_r1_READ1.fq
DRR049557_BC_W12E_r1_READ2.fq
DRR049557_BC_W12E_r2_READ1.fq
DRR049557_BC_W12E_r2_READ2.fq
DRR049557_BC_c33a_r1_READ1.fq
DRR049557_BC_c33a_r1_READ2.fq
DRR049557_BC_c33a_r2_READ1.fq
DRR049557_BC_c33a_r2_READ2.fq
DRR049557_BC_c33a_r3_READ1.fq
DRR049557_BC_c33a_r3_READ2.fq
DRR049557_un_READ1.fq
DRR049557_un_READ2.fq

Example log file

The tagdust command also produces a summary file, DRR049557_logfile.txt. In our test run, its content was:

[2016-07-07 09:29:55]	Tagdust 2.33, Copyright (C) 2013 Timo Lassmann <[email protected]>
[2016-07-07 09:29:55]	cmd: tagdust -t 4 -show_finger_seq -ref tagdust.fa -arch DRR049557.arch -o DRR049557 DRR049557_1.fastq.bz2 DRR049557_2.fastq.bz2 
[2016-07-07 09:29:55]	Start Run
--------------------------------------------------
[2016-07-07 09:29:55]	Looking at file:DRR049557_1.fastq.bz2
[2016-07-07 09:29:55]	Searching for best architecture in file 'DRR049557.arch'
[2016-07-07 09:30:26]	Using: -1 B:ACATGA,ATCATA,CACGTG,CGATGA,GAGATA,GCTCTC,GTATGA,TCGATA,AGTAGC,ATCGCA,CACTCT,CTGAGC,GAGCGT,GCTGCA,TATAGC,CACGAT,CTGACG -2 F:NNNNNNNN -3 S:TATAGGG -4 R:N 
[2016-07-07 09:30:26]	1.00 Confidence.
[2016-07-07 09:30:26]	Looking at file:DRR049557_2.fastq.bz2
[2016-07-07 09:30:26]	Searching for best architecture in file 'DRR049557.arch'
[2016-07-07 09:30:53]	Using: -1 R:N 
[2016-07-07 09:30:53]	1.00 Confidence.
[2016-07-07 09:31:10]	Determining threshold for read0.
[2016-07-07 09:31:20]	Long sequence found. Need to realloc model...
[2016-07-07 09:34:31]	Selected Threshold:: 2.658036
[2016-07-07 09:34:31]	Determining threshold for read1.
[2016-07-07 09:34:41]	Long sequence found. Need to realloc model...
[2016-07-07 09:34:42]	Selected Threshold:: 3.051751
[2016-07-07 09:35:01]	Detected casava 1.8 format.
[2016-07-08 00:24:06]	Done.

[2016-07-08 00:24:06]	DRR049557_1.fastq.bz2	Input file 0.
[2016-07-08 00:24:06]	DRR049557_2.fastq.bz2	Input file 1.
[2016-07-08 00:24:06]	217752577	total input reads
[2016-07-08 00:24:06]	3.05	selected threshold
[2016-07-08 00:24:06]	187155993	successfully extracted
[2016-07-08 00:24:06]	85.9%	extracted
[2016-07-08 00:24:06]	12041093	problems with architecture
[2016-07-08 00:24:06]	1251417	barcode / UMI not found
[2016-07-08 00:24:06]	0	too short
[2016-07-08 00:24:06]	10318	low complexity
[2016-07-08 00:24:06]	17293756	match artifacts:
[2016-07-08 00:24:06]	1750534	RT_(without_random_bases)
[2016-07-08 00:24:06]	15346	empty_(TS_linker_+_RT_reverse-complemented)
[2016-07-08 00:24:06]	470963	Nextera_501
[2016-07-08 00:24:06]	108	Nextera_502
[2016-07-08 00:24:06]	1349	Nextera_503
[2016-07-08 00:24:06]	391	Nextera_504
[2016-07-08 00:24:06]	14	Nextera_505
[2016-07-08 00:24:06]	1	Nextera_506
[2016-07-08 00:24:06]	73	Nextera_507
[2016-07-08 00:24:06]	3	Nextera_508
[2016-07-08 00:24:06]	11706684	Nextera_701
[2016-07-08 00:24:06]	374315	Nextera_702
[2016-07-08 00:24:06]	2972225	Nextera_703
[2016-07-08 00:24:06]	16	Nextera_704
[2016-07-08 00:24:06]	302	Nextera_705
[2016-07-08 00:24:06]	151	Nextera_706
[2016-07-08 00:24:06]	416	Nextera_707
[2016-07-08 00:24:06]	28	Nextera_708
[2016-07-08 00:24:06]	137	Nextera_709
[2016-07-08 00:24:06]	16	Nextera_710
[2016-07-08 00:24:06]	445	Nextera_711
[2016-07-08 00:24:06]	169	Nextera_712
[2016-07-08 00:24:06]	60	Nextera_501_Reversed:
[2016-07-08 00:24:06]	2	Nextera_502_Reversed:
[2016-07-08 00:24:06]	2	Nextera_503_Reversed:
[2016-07-08 00:24:06]	3	Nextera_504_Reversed:
[2016-07-08 00:24:06]	1	Nextera_505_Reversed:
[2016-07-08 00:24:06]	2	Nextera_703_Reversed:

The resulting FASTQ reads look like this:

@DRR049557.7 HWI-ST549:177:C6Y04ACXX:2:1101:1316:1986 length=51;FP:GTCAGGGG;RQ:30.19
GGGGAATCAGGGTTCGATTCCGGAGAGGGN
+
JJDDDDDDDDDDDDDDDEDDEDDDDDDDD#

The sequence of the fingerprint is indicated by the FP: tag in the semicolon-separated part of the read name.