Questions about CODEX2

[zhangyi]

Dear professor Jiang:

 

Excuse me ! My name is Yi Zhang.  I want to analyze CNVs from whole exome sequencing data. And I have some questions about the software “CODEX2". In the part of  running CODEX2 without specifying negative control  samples/regions, I have to set chr.index. I want to analyze all chromosomes. So I want to know if I need to set chr.index. For example, 

#chr <- 20 # This can be run for one chromosome or multiple chromosomes
#chr.index <- which(seqnames(ref_qc)==chr)
normObj.null <- normalize_null(Y_qc = Y_qc,
                               gc_qc = gc_qc,
                               K = 1:5, N = N)

So  in segmentation part,  whether I need to set chr? just like 

finalcall.CBS <- segmentCBS(Y_qc,  # recommended
                            Yhat.ns, optK = which.max(BIC.ns),
                            K = 1:5,
                            sampname_qc = colnames(Y_qc),
                            ref_qc = ranges(ref_qc),
                            lmax = 400, mode = "integer")

And finally, I want to filter CNV  from the result of finalcall.CBS( is it suitable?).  How should I choose specific standards? （like lratio > ? and so on）

Thank you, and I look forward to hearing from you.

Best regards,

Yi Zhang

------------------

zhangyi

 

[Yuchao Jiang]

Hi

Thanks for your interest in CODEX2. 

No, if you want to run normalization by CODEX2 across the entire genome, you don’t need the chr.index argument. However, for segmentation, since it applies to each chromosome separately, you will need to provide with the chr.index argument.

For targeted sequencing though, since the targets are very sparse, segmentation is performed within each gene. See here for more details https://github.com/yuchaojiang/CODEX2#codex2-for-targeted-sequencing

Hope that this helps with clarification. Please let us know if you have further questions.

Yuchao

--
You received this message because you are subscribed to the Google Groups "CODEX/CODEX2 for CNV detection" group.
To unsubscribe from this group and stop receiving emails from it, send an email to codex2+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/codex2/tencent_03CC9D459F4E82FF7CF1873784BFD9181006%40qq.com.

[Yuchao Jiang]

mBIC is returned as one of the output values from segmentCBS.

If you want to call longer CNVs, you need to set the argument lmax to be larger.

On Wed, Sep 18, 2019 at 3:58 AM zhangyi <10452...@qq.com> wrote:

Dear professor Jiang:

In the previous e-mail, you recommand filtering based on mBIC values. As far as I know, the lowest mBIC is preferred. I don't know whether it is right. I can not find a suitable mBIC to filter. And in the manual of codex2, CNV length  greater than 200kb will be failed.  Is codex2 preferred small CNV？

Thank you !

Best wishes !

------------------

zhangyi

 

------------------ 原始邮件 ------------------

发件人: "Yuchao Jiang"<yj...@cornell.edu>;

发送时间: 2019年9月4日(星期三) 上午10:09

收件人: "zhangyi"<10452...@qq.com>;

抄送: "Yuchao Jiang"<yj...@cornell.edu>;

主题: Re: [codex_wes_cnv] Questions about CODEX2

Hi Yi Zhang,

I would recommend filtering based on mBIC values (which is likelihood ratios minus a penally term). Our practical filtering strategies include CNV lengths (both bp and number of exons). Also, you should load your segmentation results into IGV for visualization. That usually gives you a goof feeling on how to proceed.

Yuchao

On Sep 3, 2019, at 9:56 PM, zhangyi <10452...@qq.com> wrote:

Dear professor Jiang:

 I'm sorry to bother you again. I still have a problem again.  I have applied segmentCBS to each chromsome separately in whole exome sequencing data. But I got too many CNVs. Before frequency filtering, could I have a preliminary screening based on these 7 columns(  length_kb  ; st_exon ; ed_exon ;raw_cov ; norm_cov ; copy_no; lratio ;mBIC)?  If it is possible , could you please tell me the detailed filter standard? 

Thank you, and I look forward to hearing from you.

Best wishes!

Yi Zhang

------------------

zhangyi

 

------------------ 原始邮件 ------------------

发件人: "Yuchao Jiang"<yj...@cornell.edu>;

发送时间: 2019年9月1日(星期天) 晚上10:59

收件人: "zhangyi"<10452...@qq.com>;"CODEX/CODEX2 for CNV detection"<cod...@googlegroups.com>;

抄送: "Yuchao Jiang"<yuc...@ad.unc.edu>;

主题: Re: [codex_wes_cnv] Questions about CODEX2