Copy number variation..on chromosome level...or ploidy with sequencing

[antu82]

I was wondering if we can use low coverage sequencing to detect the copy number variation. I am not interested in gene level or micro level but rather a whole chromosome duplication, deletion or even the whole ploidy change. I know flow cytometry is cheaper option to determine ploidy but wondering if it is achievable with multiplexing in a single lane of Illumina Hiseq. And oh, I do not have reference sequences.

Thank you

[stefanoberri]

hi antu82,

I would suggest you to give a look at CNAnorm (here the paper that describe it) We developed it to analyse low coverage data and we usually run up to 20 samples per GAIIX lane using tags. A milion mapped reads give you a decent resolution (comparable to aCGH)

As a reference, we usually use a pool of reads from the 1000 genome project, divided by gender (a pool of males and a pool of females).

if the tumour is not heterogeneous, you should also be able to get the absolute ploidy, especially if you have a rought estimate of tumour content.

[antu82]

stefanoberi,

Thanks for the response.
I looked at the paper and the woods et al (2010) you cited was one of the paper I reviewed earlier and came to this thought process. However, In the methods section, I see the PCR and am concerned if PCR affects the read number.
As you are using GAIIx and multiplex upto 20 samples, I am hopeful that I can go at least 3 times more sample in hiseq given 5-6 times number of reads in hiseq.
I am also not working on human rather a plant species with no reference genome.


Thank you for some insight.

[HESmith]

In the absence of a reference for alignment, it's unclear to me how low-coverage sequencing will provide information regarding copy number variation. Assuming your samples are all from the same species, you could generate a de novo assembly from the entire data set, then align the individual samples to your contigs to determine average read depth as an indicator of copy number.

[stefanoberri]

However, In the methods section, I see the PCR and am concerned if PCR affects the read number.

The PCR increases the total number of DNA fragments, but you do not sequence the same molecule twice (because you sequence a very small subset of the available ones). If you still are worried, you could remove PCR duplicates (samtools and picard do that)

I am also not working on human rather a plant species with no reference genome.

I agree with HESmith that this is a major problem. The system relies on the fact that you can tell where reads come from because you have a reference genome. What do you plan to do with your fastq files with the reads?

[antu82]

Thanks HESmith and stefanoberri,

I was hoping that even in the absence of reference genome, if I sequence multiple genotypes, I can compare to each other for relative copy numbers. Probably I may have access to just contig-level-assembly reference.

Thanks

[SeqAK]

Hi,
One can reduce the complexity in genome by using methylation sensitive REs and size selection during library construction for sequencing. Later read depth algorithms could be used for estimating copy number variations. This might be helpful http://jxb.oxfordjournals.org/conten...rs105.abstract