visualization of output of QuEST

[jlmlj]

Hi Anton!

I have a question regarding to visualization of output of QuEST 2.4
using UCSC genome browser.

Basically I was trying to upload a custom tract following your
instruction:
“when the File Upload menu comes up, navigate to QuEST_analysis/tracks/
wig_profiles/by_chr/ChIP_unnormalized" directory and select
"chr11.wig.gz". Click on "Open" button of the menu. Then click on the
"Submit" button of the genome browser.”

First, the QuEST generated .wig files with names like “11.wig.gz” on
the computer. When I uploaded as a custom tract, it always returned me
error massages: “Error File '11.wig.gz' - track load error (track
name='ct_ChIPnormalized11_4560'):
Couldn't find size of chromosome 11 (note: chrom names are case
sensitive)” whenever I change the name into “chr11.wig.gz” or not. Do
you know why?

Moreover, I saw one of your posts saying that you used “normalized”
file rather than “unnormalized”, which was different from your
instruction. Would you mind mentioning the reason?

Many thanks for help!

[Hunter Richards]

regarding the upload error, did you make sure you have the right
reference genome release selected? You'll get errors like that if you
try to load a track from reads aligned to say the mouse genome to a
human reference genome on UCSC or from the same species but to a
different release than which was used in the alignment (e.g. mm9 aligned
- mm7 on UCSC).

Hunter Richards
Postdoctoral Scientist
Life Sciences Division
Lawrence Berkeley National Laboratory
1 Cyclotron Rd, Bldg 84-155
University of California, Berkeley, 94720
510-486-4663

jlmlj wrote:
> Hi Anton!
>
> I have a question regarding to visualization of output of QuEST 2.4
> using UCSC genome browser.
>
> Basically I was trying to upload a custom tract following your
> instruction:

> �when the File Upload menu comes up, navigate to QuEST_analysis/tracks/

> wig_profiles/by_chr/ChIP_unnormalized" directory and select
> "chr11.wig.gz". Click on "Open" button of the menu. Then click on the

> "Submit" button of the genome browser.�
>
> First, the QuEST generated .wig files with names like �11.wig.gz� on

> the computer. When I uploaded as a custom tract, it always returned me

> error massages: �Error File '11.wig.gz' - track load error (track

> name='ct_ChIPnormalized11_4560'):
> Couldn't find size of chromosome 11 (note: chrom names are case

> sensitive)� whenever I change the name into �chr11.wig.gz� or not. Do
> you know why?
>
> Moreover, I saw one of your posts saying that you used �normalized�
> file rather than �unnormalized�, which was different from your

[jlmlj]

Thanks a lot, Hunter!
I tried again to make sure I used hg18, however I still have such an
error...
btw, the reads were aligned to hg18 using bowtie in my case.

On Jul 1, 7:35 pm, Hunter Richards <hwricha...@lbl.gov> wrote:
> regarding the upload error, did you make sure you have the right
> reference genome release selected? You'll get errors like that if you
> try to load a track from reads aligned to say the mouse genome to a
> human reference genome on UCSC or from the same species but to a
> different release than which was used in the alignment (e.g. mm9 aligned
> - mm7 on UCSC).
>
> Hunter Richards
> Postdoctoral Scientist
> Life Sciences Division
> Lawrence Berkeley National Laboratory
> 1 Cyclotron Rd, Bldg 84-155
> University of California, Berkeley, 94720
> 510-486-4663
>
> jlmlj wrote:
> > Hi Anton!
>
> > I have a question regarding to visualization of output of QuEST 2.4
> > using UCSC genome browser.
>
> > Basically I was trying to upload a custom tract following your
> > instruction:

> > when the File Upload menu comes up, navigate to QuEST_analysis/tracks/
> > wig_profiles/by_chr/ChIP_unnormalized" directory and select
> > "chr11.wig.gz". Click on "Open" button of the menu. Then click on the
> > "Submit" button of the genome browser.
>

> > First, the QuEST generated .wig files with names like 11.wig.gz on

> > the computer. When I uploaded as a custom tract, it always returned me

> > error massages: Error File '11.wig.gz' - track load error (track

> > name='ct_ChIPnormalized11_4560'):
> > Couldn't find size of chromosome 11 (note: chrom names are case

> > sensitive) whenever I change the name into chr11.wig.gz or not. Do
> > you know why?
>
> > Moreover, I saw one of your posts saying that you used normalized
> > file rather than unnormalized , which was different from your

[Noboru Jo Sakabe]

You say that the wig files generated are named 11.wig.gz. They
should be named chr11.wig.gz
To me it seems that when you aligned with Bowtie, the chromosomes
were not names as UCSC likes.
So you will have to edit your .wig files and rename the chromosomes
to the UCSC convention. Or realign the reads.
Hope this helps.

[jlmlj]

Thanks for the response, Noboru!
I changed the name but it did not work too...

>  nsakabe.vcf
> < 1KViewDownload

[jlmlj]

Noboru,

I meant I renamed the file name from '11.wig.zg' into 'chr11.wig.gz',
and it did not work.

I am not sure if it is what your said about "rename the chromosomes to
the UCSC convention"?

Thanks!

[Noboru Jo Sakabe]

The name of the file doesn't matter. I just mentioned the name
because QuEST probably names the file after the chromosome names in your
alignment files. That gave me the hint that you have the wrong
chromosome names (non-UCSC) in your alignments.
So either you realign (which is silly) to a Bowtie index with UCSC
chromosome names (chr11 instead of 11) or edit the wig's.
You have to edit the names of the chromosomes inside the file.
The error that you got from UCSC says it can't find the size for
"chromosome 11", because it was looking for "chromosome chr11".
If you read the contents of your wig file there will be a header
saying something like:

track type=wiggle_0 name="TF_normalized_chr11"
description="TF_normalized_chr11" visibility=full color=227,92,92
priority=63 maxHeightPixels=50:50:11 variableStep chrom=11 span=1

You have to change chrom=11 to chrom=chr11.

track type=wiggle_0 name="TF_normalized_chr11"
description="TF_normalized_chr11" visibility=full color=227,92,92
priority=63 maxHeightPixels=50:50:11 variableStep chrom=chr11 span=1

[jlmlj]

Thank you very much, Noboru! It works well now~!

>  nsakabe.vcf
> < 1KViewDownload

[jlmlj]

Noboru,
btw, I am thinking of realigning the reads because I will have many
samples later on. I am curious how it happened because I followed the
instruction of Bowtie carefully. Do you have a hint?
Thanks again!

[Noboru Jo Sakabe]

It's been a long time since I obtained my bowtie indexes, but it
seems that I downloaded this file that contained all the indexes.
ftp://ftp.cbcb.umd.edu/pub/data/bowtie_indexes/m_musculus_ncbi37.ebwt.zip

I just tested and it gave me UCSC chromosome names.

When I look for chromosome names in my index files, I do get UCSC
names (below is my output). Try this command and see what you get:

strings mm9.1.ebwt | grep chr

Otherwise, I don't know why you're getting different names.

$ strings mm9.1.ebwt | grep chr
chrE
chrJ*
vSchr
Fchr
uVchr
chr1
chr2
chr3
chr4
chr5
chr6
chr7
chr8
chr9
chr10
chr11
chr12
chr13
chr14
chr15
chr16
chr17
chr18
chr19
chrX
chrY
chrM

[jlmlj]

Hi Noboru,

As to bowtie indexes, I download the hg18 and built my own indexes. It
seems I do not have UCSC names. What I have is as below(one of 6
indexes):
-system-specific-3.2$ strings h_sapiens_asm.1.ebwt | grep chr
MchrT7
Tachrj
chr`
chrR
chr)D
Echr
chr=O

Do you have any idea about this?
Thanks!

On Jul 2, 3:43 pm, Noboru Jo Sakabe <nsak...@uchicago.edu> wrote:
>     It's been a long time since I obtained my bowtie indexes, but it

> seems that I downloaded this file that contained all the indexes.ftp://ftp.cbcb.umd.edu/pub/data/bowtie_indexes/m_musculus_ncbi37.ebwt...

>  nsakabe.vcf
> < 1KViewDownload

[Noboru Jo Sakabe]

Did you download from UCSC? Maybe you got from Ensembl?
Why don't you use the indexes from the Bowtie website?

ftp://ftp.cbcb.umd.edu/pub/data/bowtie_indexes/hg18.ebwt.zip

[Hunter Richards]

I think what was meant was that the indexes used to align your data used
the NCBI nomenclature rather than that from UCSC. I think NCBI names
chromosomes based on accession naming, or something like that. I think I
read that somewhere. If you look on the bowtie website (left hand side
of the page) you'll see pre-built indexes with entries for the NCBI and
UCSC versions http://bowtie-bio.sourceforge.net/index.shtml

[Anton]

There is also a danger of displaying UCSC tracks based on the sequence
that you obtained from elsewhere - the coordinates may not match
correctly.

[jlmlj]

Thanks, everyone! your suggestions help me a lot!
The reason I did not use the index provided online was because I
thought it could be slow. Now I understand I used NCBI index rather
than UCSC one. So I will try the pre-indexed one.

Thanks again!

[luca magnani]

Dear all,

can you reccomend a terminal window command to reformat wig files from chr=1 to chr=chr1?

I have tried a SED based approach but I may be doing something stupid (very limited with coding).

Thanks!

Luca

[Hunter Richards]

$ sed 's/chr=1/chr=chr1/g' file-in > file-out

"s" means substitute

The string in the first part of the expression "/chr=1/" is what you are searching for in your document

The second part "chr=chr1/" is what you want to replace the search string with

"g" means globally. Otherwise it would replace just the first instance of "chr=1".

Here's a few handy guides for sed commands.

http://www.catonmat.net/blog/sed-one-liners-explained-part-one/
http://www.grymoire.com/Unix/Sed.html

http://www.cyberciti.biz/tips/delete-leading-spaces-from-front-of-each-word.html

Take some time to learn them. It'll be super handy with these research tools. Also learn awk. The two together are a powerful way of manipulating text files at the command line.

cheers,

-hunter

--
 
---
You received this message because you are subscribed to the Google Groups "QuEST ChIP-seq group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chipseq+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
Hunter Richards
Postdoctoral Scientist
Department of Genome Dynamics
Lawrence Berkeley National Laboratory
1 Cyclotron Rd, Bldg 977-138