Cnv Analysis On Genes

Question

[Adrian Pelin]

Hello,

Is there a CNV variant caller that works on non-model organisms and that works by providing a bam alignment, a fasta reference and a gff annotation of all genes?

Basically I am trying to find our what genes are in high copy numbers.

I am creating the bam file by aligning illumina paired end reads with bwa to reference.

Thank you, Adrian

Answer 1

[Charles Warden]

Depends on your application, but the input files should typically be the same (for common and non-model organisms).

I happen to like CoNIFER if you are doing exon-capture analysis:

http://conifer.sourceforge.net/

You can also use DNAcopy to call segments based upon the normalized values from CoNIFER (and can theoretically work with any table of normalized abundances across windows in the genome):

http://www.bioconductor.org/packages/2.13/bioc/html/DNAcopy.html

[Charles Warden]

Also, VarScan can technically work on any .pileup file produced from a .bam files, but I wouldn't trust the copy number calls unless you had paired samples (and used the "copynumber" function instead of the "mpileup2cns"). And you'll need to use another tool to get gene annotations.

[Adrian Pelin]

my organism is intronless, so conifer seems inappropriate. VarScan doesn't care about gene annotations, and reports variation. I just need to know what is the copy number in the genome of each gene I feed the program from my annotation file.

[Charles Warden]

Although designed for being given a list of targeted regions, you can define your "targeted" regions as your genes. CoNIFER doesn't actually know or care what the regions represent.

That said, I can't absolutely say this is the best strategy - it is just what I can think of.

[Denis]

I use cn.mops package (http://www.bioconductor.org/packages/release/bioc/html/cn.mops.html) exactly for the same purposes. We have reference, a number of BAM files and also GFF file. If interested i could provide an example of the R code for the analysis.

[Lei]

Can you share your R code? Thanks!