DEGseq

[RockChalkJayhawk]

Hey all,

A new package has come out for RNA Seq Analysis:

DEGseq: an R package for identifying differentially expressed genes from RNA-seq data.
Wang L, Feng Z, Wang X, Wang X, Zhang X.
Bioinformatics. 2009 Oct 24. [Epub ahead of print]
PMID: 19855105 [PubMed - as supplied by publisher]

However, from my understanding (which may not be correct) they only use tags that map to the genome, not splice juntions. Does anyone else see this and if so, what would be the work-around?

[svl]

I seem to be unable to install the package...anyone had succes?
----
source("http://bioconductor.org/biocLite.R")
biocLite("DEGseq")
----

Also their site is unavailable: http://bioinfo.au.tsinghua.edu.cn/software/degseq
Here is the page at Bioc: http://www.bioconductor.org/packages...ml/DEGseq.html

[RockChalkJayhawk]

I seem to be unable to install the package...anyone had succes?
----
source("http://bioconductor.org/biocLite.R")
biocLite("DEGseq")
----

Also their site is unavailable: http://bioinfo.au.tsinghua.edu.cn/software/degseq
Here is the page at Bioc: http://www.bioconductor.org/packages...ml/DEGseq.html

I had to try it a few times before I could get it loaded. Good luck.

[svl]

Thanks. I figured it out. Had to update R and Bioconductor before trying to install the package....should keep my software better updated I guess

[Xi Wang]

hi svl,

Thanks for using DEGseq.

If you want to install DEGseq from bioconductor, you should update your R to v2.10.0. Or you can install DEGseq by typing the following.

--
source("http://bioinfo.au.tsinghua.edu.cn/software/degseq/DEGseqInstall.R")
--

Best regards,
Xi

I seem to be unable to install the package...anyone had succes?
----
source("http://bioconductor.org/biocLite.R")
biocLite("DEGseq")
----

Also their site is unavailable: http://bioinfo.au.tsinghua.edu.cn/software/degseq
Here is the page at Bioc: http://www.bioconductor.org/packages...ml/DEGseq.html

[Xi Wang]

Sorry that yesterday our building encountered a power cut, so our web server was not at work. Our site works well now.

Thanks.
Xi

Also their site is unavailable: http://bioinfo.au.tsinghua.edu.cn/software/degseq
Here is the page at Bioc: http://www.bioconductor.org/packages...ml/DEGseq.html

[Xi Wang]

Hi RockChalkJayhawk,

Thanks for using DEGseq.

In the current version of DEGseq, we do not consider the splice junctions, and so do the reads with multiple matches to the reference genome. Actually, we take the reads mapped to the reference genome not transcriptiome as input, and count the reads in the annotateed gene regions as the gene expressing level. It works if the gene expression patterns (isoform expression percentage) are similar between the case and control samples, although otherwise it may cause a litter bias. We are now working on how to use the information provided by splice junction reads and multiple aligned reads to refine the work on differetially expressed gene identification.

Xi

Hey all,

A new package has come out for RNA Seq Analysis:

DEGseq: an R package for identifying differentially expressed genes from RNA-seq data.
Wang L, Feng Z, Wang X, Wang X, Zhang X.
Bioinformatics. 2009 Oct 24. [Epub ahead of print]
PMID: 19855105 [PubMed - as supplied by publisher]

However, from my understanding (which may not be correct) they only use tags that map to the genome, not splice juntions. Does anyone else see this and if so, what would be the work-around?

[Xi Wang]

Thanks. I figured it out. Had to update R and Bioconductor before trying to install the package....should keep my software better updated I guess

Sorry for the inconvenience.

If you want to install DEGseq through bioconductor by the followsing script, you need to update your R to 2.10.0 and bioc to 2.5

source("http://bioconductor.org/biocLite.R")
biocLite("DEGseq")

Or alternatively, you can install DEGseq through our site by:

source("http://bioinfo.au.tsinghua.edu.cn/software/degseq/DEGseqInstall.R")

Yesterday, our building encountered a power cut, so the server was down. The server is running well now.

Thanks and Best wishes,
Xi

[svl]

Sorry for the inconvenience.

Thanks for the update/info!

[shibujohn]

Is it possible to analyze SOLiD rna-seq in DEGseq

Hi RockChalkJayhawk,

Thanks for using DEGseq.

In the current version of DEGseq, we do not consider the splice junctions, and so do the reads with multiple matches to the reference genome. Actually, we take the reads mapped to the reference genome not transcriptiome as input, and count the reads in the annotateed gene regions as the gene expressing level. It works if the gene expression patterns (isoform expression percentage) are similar between the case and control samples, although otherwise it may cause a litter bias. We are now working on how to use the information provided by splice junction reads and multiple aligned reads to refine the work on differetially expressed gene identification.

Xi

Hi,
Is it possible to analyze SOLiD rna-seq data in DEGseq?

Shibu

[Xi Wang]

Hi,
Is it possible to analyze SOLiD rna-seq data in DEGseq?

Shibu

Thanks for your question.

We only used the assumption that reads are uniformly distributed along transcripts. We can upload your uniquely mapped reads (or profiles) to a browser (such as UCSC genome browser) to check whether your data satisfy this assumption. And less stringently, you can check whether the variation between technical replicates (if any) can be explained by the random sampling model. A feature of our DEGseq package can help check it: see Section 3 in Supplementary Material on line for details. May this information helps you.

Xi

[shibujohn]

SOLiD WT-data

Thanks for your question.

We only used the assumption that reads are uniformly distributed along transcripts. We can upload your uniquely mapped reads (or profiles) to a browser (such as UCSC genome browser) to check whether your data satisfy this assumption. And less stringently, you can check whether the variation between technical replicates (if any) can be explained by the random sampling model. A feature of our DEGseq package can help check it: see Section 3 in Supplementary Material on line for details. May this information helps you.

Xi

Yes, We have the SOLiD data which is uniquely mapped into the refseq (two groups BN & SHR).. And I don't have the replicates.. We had performed the Whole transcriptome in SOLiD, so is it possible to analyze the SOLiD uniquely mapped reads (.ma)?
Thanks,
Shibu

[Xi Wang]

Yes, We have the SOLiD data which is uniquely mapped into the refseq (two groups BN & SHR).. And I don't have the replicates.. We had performed the Whole transcriptome in SOLiD, so is it possible to analyze the SOLiD uniquely mapped reads (.ma)?
Thanks,
Shibu

It looks possible. You'd better also try to valid the results given by DEGseq. Good luck!

Xi

[adamreid]

Hi there,

When using DEGexp with read counts for each gene, how are the read counts normalised? I've put dummy values in for gene lengths. Does it use these, expecting
no exons?

Adam

[Xi Wang]

Hi there,

When using DEGexp with read counts for each gene, how are the read counts normalised? I've put dummy values in for gene lengths. Does it use these, expecting
no exons?

Adam

We don't recommend to normalize the read counts, although you can use RPKM as the gene expression level. DEGexp considers the number of total reads that map to the gene exon regions, when the sequence depth between samples are not the same. Why we do like this is to make sure the computation under the assumption of the random sampling model.

I am not sure I understand quite well what you meant by asking the second question. When using DEGexp, the gene expression levels are provided by your data. You can define the "gene expression" yourself. However, we recommend that you use the raw read counts as the gene expression levels.

Thanks for your question.
Xi