dwgsim usage

[gprakhar]

I want to use dwgsim to generate CNVs and InDels.
Searched for documentation, didn't find anything useful except "http://sourceforge.net/apps/mediawiki/dnaa/index.php?title=Whole_Genome_Simulation".

Tried using it myself, its working, am getting an output, hence have some questions!

1) "chr1 763 A C -" what does this line of output mean?
As far as I can understand the program is taking Nucleotides from Ref Chr1, and changing them based on some parameter and then generating reads.
In short can you please explain the algorithm behind this??

In the output format, what is meant by "start end 1" & "start end 2" ?

2) Why does the program produce 3 output files(2 for bwa, 1 for BFast)?

Any help will be highly appreciated!!!

[nilshomer]

I want to use dwgsim to generate CNVs and InDels.
Searched for documentation, didn't find anything useful except "http://sourceforge.net/apps/mediawiki/dnaa/index.php?title=Whole_Genome_Simulation".

Tried using it myself, its working, am getting an output, hence have some questions!

The source code is the best documentation. This simulator is forked off of the SAMtools simulator, which was modified from the MAQ simulator. It does not simulate CNVs, but only small indels, SNPs, and a single type of sequencing error (base change).

1) "chr1 763 A C -" what does this line of output mean?
As far as I can understand the program is taking Nucleotides from Ref Chr1, and changing them based on some parameter and then generating reads.
In short can you please explain the algorithm behind this??

It randomly adds SNPs and inserts/deletes bases at the given rates (-r/-R) in the given reference. The indel length is determined by an exponential distribution (-X). Then reads are simulated from this mutated genome with the given error rate (-e). Paired end reads' insert size are also drawn from an distribution. See the source code for more details.

In the output format, what is meant by "start end 1" & "start end 2" ?

The start positions in the reference from where the two paired end reads were drawn.

2) Why does the program produce 3 output files(2 for bwa, 1 for BFast)?

Any help will be highly appreciated!!!

So we can test BWA and BFAST on the same datasets. BWA requires two files, one for each paired end, while BFAST requires an aggregated file.

[gprakhar]

Thank you for the help.
I do agree that Source Code is the best documentation, still I will try and create a manual, as I learn about the tool.
We have a CNV detection tool which we want to test. Hence I wanted simulated data with CNV at known locations.

Any simulators for this particular job?

SV simulation from the PEMer pacakage looks good, but I am having trouble compiling it.

[Alex8]

Hello, Nils,
I have a problem obtaining color-space simulation reads.
The output BFAST looks OK:

@gi|12057207|gb|AE001439.1|_637443_636958_1_1_1:0:0_0:0:0_0
A21322003132201120203003022220220001033010212331013
+
11111111111111111111111111111111111111111111111111
...

But the two BWA files contain nucleotide reads:

@gi|12057207|gb|AE001439.1|_637444_636959_1_1_1:0:0_0:0:0_0/2
CTGGAATCTGGACCGAGATAATAGGGGAGGAAACATTACAGCGTTCACT
+
1111111111111111111111111111111111111111111111111
...

Is is intended? How to get BWA files in color-space?

Regards,
Alexander

[nilshomer]

Hello, Nils,
I have a problem obtaining color-space simulation reads.
The output BFAST looks OK:

@gi|12057207|gb|AE001439.1|_637443_636958_1_1_1:0:0_0:0:0_0
A21322003132201120203003022220220001033010212331013
+
11111111111111111111111111111111111111111111111111
...

But the two BWA files contain nucleotide reads:

@gi|12057207|gb|AE001439.1|_637444_636959_1_1_1:0:0_0:0:0_0/2
CTGGAATCTGGACCGAGATAATAGGGGAGGAAACATTACAGCGTTCACT
+
1111111111111111111111111111111111111111111111111
...

Is is intended? How to get BWA files in color-space?

Regards,
Alexander

This is intended. The BWA reads are "double-encoded" and the input to BWA as if you had used BWA's solid2fastq.pl.

[Alex8]

I wonder what is the easy way to convert double-encoded to csfasta? SOLiD denovo tools seems to contain similar function but is is not clearly available.

[nilshomer]

I wonder what is the easy way to convert double-encoded to csfasta? SOLiD denovo tools seems to contain similar function but is is not clearly available.

You should use the BFAST output to convert the CSFATQ to CSFASTA/QUAL, not the BWA. The outputs of dwgsim were designed for BFAST and BWA.

[Alex8]

Thank you for the information!

[Alex8]

Nils, please let know the format of the base/color error rate file - the one that goes with -E key.

[nilshomer]

Nils, please let know the format of the base/color error rate file - the one that goes with -E key.

In the 0.1.1 release, you should have one value per line, with the ith value representing the error rate of the ith base.

The latest GIT, which is what I would like to support, does not use an error file, but instead uses two command line options "-e/-E" giving either a uniform error rate (ex "-e 0.01") or a increasing error rate (ex from 0.01 to 0.1 is "-e 0.01-0.1").

The source code is your best documentation for this software.

[arkal]

Hey, I'm using dwgsim to generate paired end reads for my data. I need to generate a few single end read files as well is there any way i can use dwgsim to do the same? I don't want to use 2 different simulators for my studies.

-Arjun

[nilshomer]

Sure, specify zero length reads for the second end ("-2 0").

[arkal]

Sure, specify zero length reads for the second end ("-2 0").

Wanted to clarify that! Thanks a ton!

-A

[arkal]

Hey i just got another doubt regarding my single end problem... if i'm setting "-2 0" then do i double the value of -N since N is the No of Read Pairs?

-A

[arkal]

Hey i just got another doubt regarding my single end problem... if i'm setting "-2 0" then do i double the value of -N since N is the No of Read Pairs?

-A

Just to clarify what i mean, supposing i need 1,000,000 read pairs for 15x coverage,
PE
-N 1000000 -1 76 -2 76

then is SE

-N 1000000 -1 76 -2 0

or

-N 2000000 -1 76 -2 0

?

Also, if I take file 1 of PE 30X is it equivalent to SE 15X?
-A