Search results are wrong. Please help.

classic Classic list List threaded Threaded
12 messages Options
Reply | Threaded
Open this post in threaded view
|

Search results are wrong. Please help.

wizi
Hi everyone, 

I am trying to search with Alias name but the results are always less than its supposed to be. The GFF is loaded to mySQL using seqfeature_load. When I search for S3_10_23, it returns 5 regions instead of 8. It also happened with other type (Alias) as well. 
Please look at the screenshot below you will understand what I mean. 



When I view S3_10_23 under Details view, it appears 8 regions. 


Please help!!!

Thank you very much, 


------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

wizi
Oh I am using Gbrowse2.13. 

My conf: 
db_adaptor    = Bio::DB::SeqFeature::Store
db_args       = -adaptor DBI::mysql
                -dsn     dbi:mysql:database=hiv_0815
                -user    xxxx
-pass xxxxx

On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
Hi everyone, 

I am trying to search with Alias name but the results are always less than its supposed to be. The GFF is loaded to mySQL using seqfeature_load. When I search for S3_10_23, it returns 5 regions instead of 8. It also happened with other type (Alias) as well. 
Please look at the screenshot below you will understand what I mean. 



When I view S3_10_23 under Details view, it appears 8 regions. 


Please help!!!

Thank you very much, 



------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

Scott Cain
In reply to this post by wizi
Hi Minh,

I think what is happening here is that, when features of the same type
share common boundaries, to the search tool, they look like the same
result, particularly if the have the same aliases.  In fact, it seems
to me that if that is the case, maybe the should all collapse into one
feature anyway.  Is there a reason they aren't?

Scott


On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:

> Hi everyone,
> I am trying to search with Alias name but the results are always less than
> its supposed to be. The GFF is loaded to mySQL using seqfeature_load. When I
> search for S3_10_23, it returns 5 regions instead of 8. It also happened
> with other type (Alias) as well.
> Please look at the screenshot below you will understand what I mean.
> http://img844.imageshack.us/img844/8972/screenshotdx.png
>
> When I view S3_10_23 under Details view, it appears 8 regions.
> http://img200.imageshack.us/img200/8070/screenshot1rr.png
> Please help!!!
> Thank you very much,
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
> _______________________________________________
> Gmod-gbrowse mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
>
>



--
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research

------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

wizi
Hi Scott, 
Could you explain more on collapsing into one feature?

Basically, what I want to know is the location of HIV clusters based on type. 

The clusters' names follow this format: GInumber-Type (Example: GI######, Type: S3_10_23) 

I want to be able to search all the location of a HIV-type. 
How should I configure Gbrowse2 to do that?

Sorry about my english. 
Let me know if you need more info. 

Minh Bui. 


On Mon, Aug 16, 2010 at 11:44 AM, Scott Cain <[hidden email]> wrote:
Hi Minh,

I think what is happening here is that, when features of the same type
share common boundaries, to the search tool, they look like the same
result, particularly if the have the same aliases.  In fact, it seems
to me that if that is the case, maybe the should all collapse into one
feature anyway.  Is there a reason they aren't?

Scott


On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
> Hi everyone,
> I am trying to search with Alias name but the results are always less than
> its supposed to be. The GFF is loaded to mySQL using seqfeature_load. When I
> search for S3_10_23, it returns 5 regions instead of 8. It also happened
> with other type (Alias) as well.
> Please look at the screenshot below you will understand what I mean.
> http://img844.imageshack.us/img844/8972/screenshotdx.png
>
> When I view S3_10_23 under Details view, it appears 8 regions.
> http://img200.imageshack.us/img200/8070/screenshot1rr.png
> Please help!!!
> Thank you very much,
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
> _______________________________________________
> Gmod-gbrowse mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
>
>



--
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research


------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

Scott Cain
Hi Minh,

So each of these features that share the same boundaries are different
sequence variants?  I'm guessing that based on the sequence notes in
the screenshot; the second and there sequences look nearly identical
(I can see at least one snp though).

I'm not sure how I would handle this.  On one hand, the search results
are correct: it says the following 5 regions (not features) match, and
if a user follows any one of those links, she will see all of the
relevant features in the display.  I understand that it may mislead
the user into thinking there are only 5 features when there are
actually eight.

To collapse the ones that share boundaries into one feature, you could
swap the Name and Alias tags, where all of the features of a given
type share the name of the type, and then the Alias would be populated
with a list of all of the names of the individual GInumber-Type
entries.  This might be less than ideal if you want to capture full
sequence for each of the variants though.

Scott


On Mon, Aug 16, 2010 at 12:19 PM, Minh Bui <[hidden email]> wrote:

> Hi Scott,
> Could you explain more on collapsing into one feature?
> Basically, what I want to know is the location of HIV clusters based on
> type.
> The clusters' names follow this format: GInumber-Type (Example: GI######,
> Type: S3_10_23)
> I want to be able to search all the location of a HIV-type.
> How should I configure Gbrowse2 to do that?
> Sorry about my english.
> Let me know if you need more info.
> Minh Bui.
>
> On Mon, Aug 16, 2010 at 11:44 AM, Scott Cain <[hidden email]> wrote:
>>
>> Hi Minh,
>>
>> I think what is happening here is that, when features of the same type
>> share common boundaries, to the search tool, they look like the same
>> result, particularly if the have the same aliases.  In fact, it seems
>> to me that if that is the case, maybe the should all collapse into one
>> feature anyway.  Is there a reason they aren't?
>>
>> Scott
>>
>>
>> On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
>> > Hi everyone,
>> > I am trying to search with Alias name but the results are always less
>> > than
>> > its supposed to be. The GFF is loaded to mySQL using seqfeature_load.
>> > When I
>> > search for S3_10_23, it returns 5 regions instead of 8. It also happened
>> > with other type (Alias) as well.
>> > Please look at the screenshot below you will understand what I mean.
>> > http://img844.imageshack.us/img844/8972/screenshotdx.png
>> >
>> > When I view S3_10_23 under Details view, it appears 8 regions.
>> > http://img200.imageshack.us/img200/8070/screenshot1rr.png
>> > Please help!!!
>> > Thank you very much,
>> >
>> >
>> > ------------------------------------------------------------------------------
>> > This SF.net email is sponsored by
>> >
>> > Make an app they can't live without
>> > Enter the BlackBerry Developer Challenge
>> > http://p.sf.net/sfu/RIM-dev2dev
>> > _______________________________________________
>> > Gmod-gbrowse mailing list
>> > [hidden email]
>> > https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
>> >
>> >
>>
>>
>>
>> --
>> ------------------------------------------------------------------------
>> Scott Cain, Ph. D.                                   scott at scottcain
>> dot net
>> GMOD Coordinator (http://gmod.org/)                     216-392-3087
>> Ontario Institute for Cancer Research
>
>



--
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research

------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

wizi
Hi Scott,
Here is how I come up with the GFF file:
Step 1: I was given about ~100,000 clusters (HIV sequences were already trimmed off so I could map these clusters to Human genome. We want to know where HIVs integrated themselves into Human Genome)
Step 2: I mapped these clusters to UCSC Human Genome using 454 GS Mapper. 
Step 3: The results were parsed to GFF using my Perl script. 

_There are 9 different types of HIV such as S3_10_23, W1_11_11. 
_My goal is to display all the locations of these types. 
 
In the note, I display the queries's sequences. And yes, they are different (they are all different on start-end positions). Some of them might have multiple hits on different chromosomes. Therefore, they might have the same "Notes" but different chromosome + start-end position . 



On Mon, Aug 16, 2010 at 12:53 PM, Scott Cain <[hidden email]> wrote:
Hi Minh,

So each of these features that share the same boundaries are different
sequence variants?  I'm guessing that based on the sequence notes in
the screenshot; the second and there sequences look nearly identical
(I can see at least one snp though).

I'm not sure how I would handle this.  On one hand, the search results
are correct: it says the following 5 regions (not features) match, and
if a user follows any one of those links, she will see all of the
relevant features in the display.  I understand that it may mislead
the user into thinking there are only 5 features when there are
actually eight.

To collapse the ones that share boundaries into one feature, you could
swap the Name and Alias tags, where all of the features of a given
type share the name of the type, and then the Alias would be populated
with a list of all of the names of the individual GInumber-Type
entries.  This might be less than ideal if you want to capture full
sequence for each of the variants though.

Scott


On Mon, Aug 16, 2010 at 12:19 PM, Minh Bui <[hidden email]> wrote:
> Hi Scott,
> Could you explain more on collapsing into one feature?
> Basically, what I want to know is the location of HIV clusters based on
> type.
> The clusters' names follow this format: GInumber-Type (Example: GI######,
> Type: S3_10_23)
> I want to be able to search all the location of a HIV-type.
> How should I configure Gbrowse2 to do that?
> Sorry about my english.
> Let me know if you need more info.
> Minh Bui.
>
> On Mon, Aug 16, 2010 at 11:44 AM, Scott Cain <[hidden email]> wrote:
>>
>> Hi Minh,
>>
>> I think what is happening here is that, when features of the same type
>> share common boundaries, to the search tool, they look like the same
>> result, particularly if the have the same aliases.  In fact, it seems
>> to me that if that is the case, maybe the should all collapse into one
>> feature anyway.  Is there a reason they aren't?
>>
>> Scott
>>
>>
>> On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
>> > Hi everyone,
>> > I am trying to search with Alias name but the results are always less
>> > than
>> > its supposed to be. The GFF is loaded to mySQL using seqfeature_load.
>> > When I
>> > search for S3_10_23, it returns 5 regions instead of 8. It also happened
>> > with other type (Alias) as well.
>> > Please look at the screenshot below you will understand what I mean.
>> > http://img844.imageshack.us/img844/8972/screenshotdx.png
>> >
>> > When I view S3_10_23 under Details view, it appears 8 regions.
>> > http://img200.imageshack.us/img200/8070/screenshot1rr.png
>> > Please help!!!
>> > Thank you very much,
>> >
>> >
>> > ------------------------------------------------------------------------------
>> > This SF.net email is sponsored by
>> >
>> > Make an app they can't live without
>> > Enter the BlackBerry Developer Challenge
>> > http://p.sf.net/sfu/RIM-dev2dev
>> > _______________________________________________
>> > Gmod-gbrowse mailing list
>> > [hidden email]
>> > https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
>> >
>> >
>>
>>
>>
>> --
>> ------------------------------------------------------------------------
>> Scott Cain, Ph. D.                                   scott at scottcain
>> dot net
>> GMOD Coordinator (http://gmod.org/)                     216-392-3087
>> Ontario Institute for Cancer Research
>
>



--
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research


------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

Lincoln Stein
In reply to this post by Scott Cain
Yes, there is a step that collapses features that share the same boundaries and types into a single feature. If you want them to appear as separate features, you need to give them a different method or source.

Lincoln

On Mon, Aug 16, 2010 at 11:44 AM, Scott Cain <[hidden email]> wrote:
Hi Minh,

I think what is happening here is that, when features of the same type
share common boundaries, to the search tool, they look like the same
result, particularly if the have the same aliases.  In fact, it seems
to me that if that is the case, maybe the should all collapse into one
feature anyway.  Is there a reason they aren't?

Scott


On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
> Hi everyone,
> I am trying to search with Alias name but the results are always less than
> its supposed to be. The GFF is loaded to mySQL using seqfeature_load. When I
> search for S3_10_23, it returns 5 regions instead of 8. It also happened
> with other type (Alias) as well.
> Please look at the screenshot below you will understand what I mean.
> http://img844.imageshack.us/img844/8972/screenshotdx.png
>
> When I view S3_10_23 under Details view, it appears 8 regions.
> http://img200.imageshack.us/img200/8070/screenshot1rr.png
> Please help!!!
> Thank you very much,
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
> _______________________________________________
> Gmod-gbrowse mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
>
>



--
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research

------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse



--
Lincoln D. Stein
Director, Informatics and Biocomputing Platform
Ontario Institute for Cancer Research
101 College St., Suite 800
Toronto, ON, Canada M5G0A3
416 673-8514
Assistant: Renata Musa <[hidden email]>

------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

Fields, Christopher J
In reply to this post by wizi
So this is akin to a seq analysis, with query/subject hits like BLAST, correct?  Have you looked at how matches and Target are defined in GFF3 (under "THE GAP ATTRIBUTE")?  You could also look into SAM/BAM and Bio-SamTools for this (which is pretty easy with GBrowse 2).

chris

On Aug 16, 2010, at 12:21 PM, Minh Bui wrote:

> Hi Scott,
> Here is how I come up with the GFF file:
> Step 1: I was given about ~100,000 clusters (HIV sequences were already trimmed off so I could map these clusters to Human genome. We want to know where HIVs integrated themselves into Human Genome)
> Step 2: I mapped these clusters to UCSC Human Genome using 454 GS Mapper.
> Step 3: The results were parsed to GFF using my Perl script.
>
> _There are 9 different types of HIV such as S3_10_23, W1_11_11.
> _My goal is to display all the locations of these types.
>  
> In the note, I display the queries's sequences. And yes, they are different (they are all different on start-end positions). Some of them might have multiple hits on different chromosomes. Therefore, they might have the same "Notes" but different chromosome + start-end position .
>
>
>
> On Mon, Aug 16, 2010 at 12:53 PM, Scott Cain <[hidden email]> wrote:
> Hi Minh,
>
> So each of these features that share the same boundaries are different
> sequence variants?  I'm guessing that based on the sequence notes in
> the screenshot; the second and there sequences look nearly identical
> (I can see at least one snp though).
>
> I'm not sure how I would handle this.  On one hand, the search results
> are correct: it says the following 5 regions (not features) match, and
> if a user follows any one of those links, she will see all of the
> relevant features in the display.  I understand that it may mislead
> the user into thinking there are only 5 features when there are
> actually eight.
>
> To collapse the ones that share boundaries into one feature, you could
> swap the Name and Alias tags, where all of the features of a given
> type share the name of the type, and then the Alias would be populated
> with a list of all of the names of the individual GInumber-Type
> entries.  This might be less than ideal if you want to capture full
> sequence for each of the variants though.
>
> Scott
>
>
> On Mon, Aug 16, 2010 at 12:19 PM, Minh Bui <[hidden email]> wrote:
> > Hi Scott,
> > Could you explain more on collapsing into one feature?
> > Basically, what I want to know is the location of HIV clusters based on
> > type.
> > The clusters' names follow this format: GInumber-Type (Example: GI######,
> > Type: S3_10_23)
> > I want to be able to search all the location of a HIV-type.
> > How should I configure Gbrowse2 to do that?
> > Sorry about my english.
> > Let me know if you need more info.
> > Minh Bui.
> >
> > On Mon, Aug 16, 2010 at 11:44 AM, Scott Cain <[hidden email]> wrote:
> >>
> >> Hi Minh,
> >>
> >> I think what is happening here is that, when features of the same type
> >> share common boundaries, to the search tool, they look like the same
> >> result, particularly if the have the same aliases.  In fact, it seems
> >> to me that if that is the case, maybe the should all collapse into one
> >> feature anyway.  Is there a reason they aren't?
> >>
> >> Scott
> >>
> >>
> >> On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
> >> > Hi everyone,
> >> > I am trying to search with Alias name but the results are always less
> >> > than
> >> > its supposed to be. The GFF is loaded to mySQL using seqfeature_load.
> >> > When I
> >> > search for S3_10_23, it returns 5 regions instead of 8. It also happened
> >> > with other type (Alias) as well.
> >> > Please look at the screenshot below you will understand what I mean.
> >> > http://img844.imageshack.us/img844/8972/screenshotdx.png
> >> >
> >> > When I view S3_10_23 under Details view, it appears 8 regions.
> >> > http://img200.imageshack.us/img200/8070/screenshot1rr.png
> >> > Please help!!!
> >> > Thank you very much,
> >> >
> >> >
> >> > ------------------------------------------------------------------------------
> >> > This SF.net email is sponsored by
> >> >
> >> > Make an app they can't live without
> >> > Enter the BlackBerry Developer Challenge
> >> > http://p.sf.net/sfu/RIM-dev2dev
> >> > _______________________________________________
> >> > Gmod-gbrowse mailing list
> >> > [hidden email]
> >> > https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
> >> >
> >> >
> >>
> >>
> >>
> >> --
> >> ------------------------------------------------------------------------
> >> Scott Cain, Ph. D.                                   scott at scottcain
> >> dot net
> >> GMOD Coordinator (http://gmod.org/)                     216-392-3087
> >> Ontario Institute for Cancer Research
> >
> >
>
>
>
> --
> ------------------------------------------------------------------------
> Scott Cain, Ph. D.                                   scott at scottcain dot net
> GMOD Coordinator (http://gmod.org/)                     216-392-3087
> Ontario Institute for Cancer Research
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev _______________________________________________
> Gmod-gbrowse mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse


------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

wizi
Hi,

To Scott:  I try to swap Alias and Name but It still gives me the same results. 
To Chris:  Yes, its like BLAST but its much faster. I mapped >90,000 clusters to UCSC Human Genome for about 15-20 minutes. BLAST or BLAT may take days.  Here is the "Hits" table that summarize how many matches for each type (only on inactivated HIVs, there is Activated HIVs which is 4 times bigger than this) 
   
UNACTIVATED
Type Match
NL3_10_21 2909
NL3_10_23 482
NL3_11_11 965
S3_10_21 6962
S3_10_23 5245
S3_11_11 3219
W3_10_21 4643
W3_10_23 2360
W3_11_11 4148
Total 30933

Let me know if you can't see it. 


To Lincoln: when you say "different method or source", you mean the 2nd column in GFF format right? Could you be more specific? I am pretty new to Gbrowse2. 

Thank you all for your help. 

Minh Bui. 



On Mon, Aug 16, 2010 at 1:45 PM, Chris Fields <[hidden email]> wrote:
So this is akin to a seq analysis, with query/subject hits like BLAST, correct?  Have you looked at how matches and Target are defined in GFF3 (under "THE GAP ATTRIBUTE")?  You could also look into SAM/BAM and Bio-SamTools for this (which is pretty easy with GBrowse 2).

chris

On Aug 16, 2010, at 12:21 PM, Minh Bui wrote:

> Hi Scott,
> Here is how I come up with the GFF file:
> Step 1: I was given about ~100,000 clusters (HIV sequences were already trimmed off so I could map these clusters to Human genome. We want to know where HIVs integrated themselves into Human Genome)
> Step 2: I mapped these clusters to UCSC Human Genome using 454 GS Mapper.
> Step 3: The results were parsed to GFF using my Perl script.
>
> _There are 9 different types of HIV such as S3_10_23, W1_11_11.
> _My goal is to display all the locations of these types.
>
> In the note, I display the queries's sequences. And yes, they are different (they are all different on start-end positions). Some of them might have multiple hits on different chromosomes. Therefore, they might have the same "Notes" but different chromosome + start-end position .
>
>
>
> On Mon, Aug 16, 2010 at 12:53 PM, Scott Cain <[hidden email]> wrote:
> Hi Minh,
>
> So each of these features that share the same boundaries are different
> sequence variants?  I'm guessing that based on the sequence notes in
> the screenshot; the second and there sequences look nearly identical
> (I can see at least one snp though).
>
> I'm not sure how I would handle this.  On one hand, the search results
> are correct: it says the following 5 regions (not features) match, and
> if a user follows any one of those links, she will see all of the
> relevant features in the display.  I understand that it may mislead
> the user into thinking there are only 5 features when there are
> actually eight.
>
> To collapse the ones that share boundaries into one feature, you could
> swap the Name and Alias tags, where all of the features of a given
> type share the name of the type, and then the Alias would be populated
> with a list of all of the names of the individual GInumber-Type
> entries.  This might be less than ideal if you want to capture full
> sequence for each of the variants though.
>
> Scott
>
>
> On Mon, Aug 16, 2010 at 12:19 PM, Minh Bui <[hidden email]> wrote:
> > Hi Scott,
> > Could you explain more on collapsing into one feature?
> > Basically, what I want to know is the location of HIV clusters based on
> > type.
> > The clusters' names follow this format: GInumber-Type (Example: GI######,
> > Type: S3_10_23)
> > I want to be able to search all the location of a HIV-type.
> > How should I configure Gbrowse2 to do that?
> > Sorry about my english.
> > Let me know if you need more info.
> > Minh Bui.
> >
> > On Mon, Aug 16, 2010 at 11:44 AM, Scott Cain <[hidden email]> wrote:
> >>
> >> Hi Minh,
> >>
> >> I think what is happening here is that, when features of the same type
> >> share common boundaries, to the search tool, they look like the same
> >> result, particularly if the have the same aliases.  In fact, it seems
> >> to me that if that is the case, maybe the should all collapse into one
> >> feature anyway.  Is there a reason they aren't?
> >>
> >> Scott
> >>
> >>
> >> On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
> >> > Hi everyone,
> >> > I am trying to search with Alias name but the results are always less
> >> > than
> >> > its supposed to be. The GFF is loaded to mySQL using seqfeature_load.
> >> > When I
> >> > search for S3_10_23, it returns 5 regions instead of 8. It also happened
> >> > with other type (Alias) as well.
> >> > Please look at the screenshot below you will understand what I mean.
> >> > http://img844.imageshack.us/img844/8972/screenshotdx.png
> >> >
> >> > When I view S3_10_23 under Details view, it appears 8 regions.
> >> > http://img200.imageshack.us/img200/8070/screenshot1rr.png
> >> > Please help!!!
> >> > Thank you very much,
> >> >
> >> >
> >> > ------------------------------------------------------------------------------
> >> > This SF.net email is sponsored by
> >> >
> >> > Make an app they can't live without
> >> > Enter the BlackBerry Developer Challenge
> >> > http://p.sf.net/sfu/RIM-dev2dev
> >> > _______________________________________________
> >> > Gmod-gbrowse mailing list
> >> > [hidden email]
> >> > https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
> >> >
> >> >
> >>
> >>
> >>
> >> --
> >> ------------------------------------------------------------------------
> >> Scott Cain, Ph. D.                                   scott at scottcain
> >> dot net
> >> GMOD Coordinator (http://gmod.org/)                     216-392-3087
> >> Ontario Institute for Cancer Research
> >
> >
>
>
>
> --
> ------------------------------------------------------------------------
> Scott Cain, Ph. D.                                   scott at scottcain dot net
> GMOD Coordinator (http://gmod.org/)                     216-392-3087
> Ontario Institute for Cancer Research
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev _______________________________________________
> Gmod-gbrowse mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse



------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

Fields, Christopher J
Minh,

My point is, if the feature is the result of an analysis where you are aligning a query to the chromosome, the use of 'Target' might be more appropriate here (in this case, the Target would be the 454 sequence along with the relevant coordinates and optional CIGAR notation).  

The alternative I also mention (SAM/BAM format): the SAMTools package comes with various starter scripts for converting BLAST, PSL, etc formats to SAM.  I use a modified one locally that adds the query sequence from a Bio::DB::Fasta to analyses for alignments.  With GBrowse2, once you have data in SAM format, convert it to a sorted BAM file (all possible using SAMTools) and add as a separate database, outlined here:

http://gmod.org/wiki/GBrowse_NGS_Tutorial

chris

On Aug 16, 2010, at 1:07 PM, Minh Bui wrote:

> Hi,
>
> To Scott:  I try to swap Alias and Name but It still gives me the same results.
> To Chris:  Yes, its like BLAST but its much faster. I mapped >90,000 clusters to UCSC Human Genome for about 15-20 minutes. BLAST or BLAT may take days.  Here is the "Hits" table that summarize how many matches for each type (only on inactivated HIVs, there is Activated HIVs which is 4 times bigger than this)
>
> UNACTIVATED
> Type Match
> NL3_10_21 2909
> NL3_10_23 482
> NL3_11_11 965
> S3_10_21 6962
> S3_10_23 5245
> S3_11_11 3219
> W3_10_21 4643
> W3_10_23 2360
> W3_11_11 4148
> Total 30933
>
> Let me know if you can't see it.
>
>
> To Lincoln: when you say "different method or source", you mean the 2nd column in GFF format right? Could you be more specific? I am pretty new to Gbrowse2.
>
> Thank you all for your help.
>
> Minh Bui.
>
>
>
> On Mon, Aug 16, 2010 at 1:45 PM, Chris Fields <[hidden email]> wrote:
> So this is akin to a seq analysis, with query/subject hits like BLAST, correct?  Have you looked at how matches and Target are defined in GFF3 (under "THE GAP ATTRIBUTE")?  You could also look into SAM/BAM and Bio-SamTools for this (which is pretty easy with GBrowse 2).
>
> chris
>
> On Aug 16, 2010, at 12:21 PM, Minh Bui wrote:
>
>> Hi Scott,
>> Here is how I come up with the GFF file:
>> Step 1: I was given about ~100,000 clusters (HIV sequences were already trimmed off so I could map these clusters to Human genome. We want to know where HIVs integrated themselves into Human Genome)
>> Step 2: I mapped these clusters to UCSC Human Genome using 454 GS Mapper.
>> Step 3: The results were parsed to GFF using my Perl script.
>>
>> _There are 9 different types of HIV such as S3_10_23, W1_11_11.
>> _My goal is to display all the locations of these types.
>>
>> In the note, I display the queries's sequences. And yes, they are different (they are all different on start-end positions). Some of them might have multiple hits on different chromosomes. Therefore, they might have the same "Notes" but different chromosome + start-end position .
>>
>>
>>
>> On Mon, Aug 16, 2010 at 12:53 PM, Scott Cain <[hidden email]> wrote:
>> Hi Minh,
>>
>> So each of these features that share the same boundaries are different
>> sequence variants?  I'm guessing that based on the sequence notes in
>> the screenshot; the second and there sequences look nearly identical
>> (I can see at least one snp though).
>>
>> I'm not sure how I would handle this.  On one hand, the search results
>> are correct: it says the following 5 regions (not features) match, and
>> if a user follows any one of those links, she will see all of the
>> relevant features in the display.  I understand that it may mislead
>> the user into thinking there are only 5 features when there are
>> actually eight.
>>
>> To collapse the ones that share boundaries into one feature, you could
>> swap the Name and Alias tags, where all of the features of a given
>> type share the name of the type, and then the Alias would be populated
>> with a list of all of the names of the individual GInumber-Type
>> entries.  This might be less than ideal if you want to capture full
>> sequence for each of the variants though.
>>
>> Scott
>>
>>
>> On Mon, Aug 16, 2010 at 12:19 PM, Minh Bui <[hidden email]> wrote:
>>> Hi Scott,
>>> Could you explain more on collapsing into one feature?
>>> Basically, what I want to know is the location of HIV clusters based on
>>> type.
>>> The clusters' names follow this format: GInumber-Type (Example: GI######,
>>> Type: S3_10_23)
>>> I want to be able to search all the location of a HIV-type.
>>> How should I configure Gbrowse2 to do that?
>>> Sorry about my english.
>>> Let me know if you need more info.
>>> Minh Bui.
>>>
>>> On Mon, Aug 16, 2010 at 11:44 AM, Scott Cain <[hidden email]> wrote:
>>>>
>>>> Hi Minh,
>>>>
>>>> I think what is happening here is that, when features of the same type
>>>> share common boundaries, to the search tool, they look like the same
>>>> result, particularly if the have the same aliases.  In fact, it seems
>>>> to me that if that is the case, maybe the should all collapse into one
>>>> feature anyway.  Is there a reason they aren't?
>>>>
>>>> Scott
>>>>
>>>>
>>>> On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
>>>>> Hi everyone,
>>>>> I am trying to search with Alias name but the results are always less
>>>>> than
>>>>> its supposed to be. The GFF is loaded to mySQL using seqfeature_load.
>>>>> When I
>>>>> search for S3_10_23, it returns 5 regions instead of 8. It also happened
>>>>> with other type (Alias) as well.
>>>>> Please look at the screenshot below you will understand what I mean.
>>>>> http://img844.imageshack.us/img844/8972/screenshotdx.png
>>>>>
>>>>> When I view S3_10_23 under Details view, it appears 8 regions.
>>>>> http://img200.imageshack.us/img200/8070/screenshot1rr.png
>>>>> Please help!!!
>>>>> Thank you very much,
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> This SF.net email is sponsored by
>>>>>
>>>>> Make an app they can't live without
>>>>> Enter the BlackBerry Developer Challenge
>>>>> http://p.sf.net/sfu/RIM-dev2dev
>>>>> _______________________________________________
>>>>> Gmod-gbrowse mailing list
>>>>> [hidden email]
>>>>> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> ------------------------------------------------------------------------
>>>> Scott Cain, Ph. D.                                   scott at scottcain
>>>> dot net
>>>> GMOD Coordinator (http://gmod.org/)                     216-392-3087
>>>> Ontario Institute for Cancer Research
>>>
>>>
>>
>>
>>
>> --
>> ------------------------------------------------------------------------
>> Scott Cain, Ph. D.                                   scott at scottcain dot net
>> GMOD Coordinator (http://gmod.org/)                     216-392-3087
>> Ontario Institute for Cancer Research
>>
>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by
>>
>> Make an app they can't live without
>> Enter the BlackBerry Developer Challenge
>> http://p.sf.net/sfu/RIM-dev2dev _______________________________________________
>> Gmod-gbrowse mailing list
>> [hidden email]
>> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev _______________________________________________
> Gmod-gbrowse mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse


------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

Scott Cain
In reply to this post by wizi
Hi Minh,

I wasn't advocating just swapping the Name and Alias, but having only one feature for the features that have the same boundaries, and putting the name values in the Alias, and putting the type in the Name.  Nevertheless, I don't think this is what you want to do if you still want to see 5 features when when there are 5 features that have the same boundaries.

I would suggest you follow the path that Lincoln suggested: assuming you aren't using the source (column 2) for anything else, you could put what is currently set as the name to the source, and then the features would be distinct when you do searches.

Scott


On Mon, Aug 16, 2010 at 2:07 PM, Minh Bui <[hidden email]> wrote:
Hi,

To Scott:  I try to swap Alias and Name but It still gives me the same results. 
To Chris:  Yes, its like BLAST but its much faster. I mapped >90,000 clusters to UCSC Human Genome for about 15-20 minutes. BLAST or BLAT may take days.  Here is the "Hits" table that summarize how many matches for each type (only on inactivated HIVs, there is Activated HIVs which is 4 times bigger than this) 
   
UNACTIVATED
Type Match
NL3_10_21 2909
NL3_10_23 482
NL3_11_11 965
S3_10_21 6962
S3_10_23 5245
S3_11_11 3219
W3_10_21 4643
W3_10_23 2360
W3_11_11 4148
Total 30933

Let me know if you can't see it. 


To Lincoln: when you say "different method or source", you mean the 2nd column in GFF format right? Could you be more specific? I am pretty new to Gbrowse2. 

Thank you all for your help. 

Minh Bui. 



On Mon, Aug 16, 2010 at 1:45 PM, Chris Fields <[hidden email]> wrote:
So this is akin to a seq analysis, with query/subject hits like BLAST, correct?  Have you looked at how matches and Target are defined in GFF3 (under "THE GAP ATTRIBUTE")?  You could also look into SAM/BAM and Bio-SamTools for this (which is pretty easy with GBrowse 2).

chris

On Aug 16, 2010, at 12:21 PM, Minh Bui wrote:

> Hi Scott,
> Here is how I come up with the GFF file:
> Step 1: I was given about ~100,000 clusters (HIV sequences were already trimmed off so I could map these clusters to Human genome. We want to know where HIVs integrated themselves into Human Genome)
> Step 2: I mapped these clusters to UCSC Human Genome using 454 GS Mapper.
> Step 3: The results were parsed to GFF using my Perl script.
>
> _There are 9 different types of HIV such as S3_10_23, W1_11_11.
> _My goal is to display all the locations of these types.
>
> In the note, I display the queries's sequences. And yes, they are different (they are all different on start-end positions). Some of them might have multiple hits on different chromosomes. Therefore, they might have the same "Notes" but different chromosome + start-end position .
>
>
>
> On Mon, Aug 16, 2010 at 12:53 PM, Scott Cain <[hidden email]> wrote:
> Hi Minh,
>
> So each of these features that share the same boundaries are different
> sequence variants?  I'm guessing that based on the sequence notes in
> the screenshot; the second and there sequences look nearly identical
> (I can see at least one snp though).
>
> I'm not sure how I would handle this.  On one hand, the search results
> are correct: it says the following 5 regions (not features) match, and
> if a user follows any one of those links, she will see all of the
> relevant features in the display.  I understand that it may mislead
> the user into thinking there are only 5 features when there are
> actually eight.
>
> To collapse the ones that share boundaries into one feature, you could
> swap the Name and Alias tags, where all of the features of a given
> type share the name of the type, and then the Alias would be populated
> with a list of all of the names of the individual GInumber-Type
> entries.  This might be less than ideal if you want to capture full
> sequence for each of the variants though.
>
> Scott
>
>
> On Mon, Aug 16, 2010 at 12:19 PM, Minh Bui <[hidden email]> wrote:
> > Hi Scott,
> > Could you explain more on collapsing into one feature?
> > Basically, what I want to know is the location of HIV clusters based on
> > type.
> > The clusters' names follow this format: GInumber-Type (Example: GI######,
> > Type: S3_10_23)
> > I want to be able to search all the location of a HIV-type.
> > How should I configure Gbrowse2 to do that?
> > Sorry about my english.
> > Let me know if you need more info.
> > Minh Bui.
> >
> > On Mon, Aug 16, 2010 at 11:44 AM, Scott Cain <[hidden email]> wrote:
> >>
> >> Hi Minh,
> >>
> >> I think what is happening here is that, when features of the same type
> >> share common boundaries, to the search tool, they look like the same
> >> result, particularly if the have the same aliases.  In fact, it seems
> >> to me that if that is the case, maybe the should all collapse into one
> >> feature anyway.  Is there a reason they aren't?
> >>
> >> Scott
> >>
> >>
> >> On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
> >> > Hi everyone,
> >> > I am trying to search with Alias name but the results are always less
> >> > than
> >> > its supposed to be. The GFF is loaded to mySQL using seqfeature_load.
> >> > When I
> >> > search for S3_10_23, it returns 5 regions instead of 8. It also happened
> >> > with other type (Alias) as well.
> >> > Please look at the screenshot below you will understand what I mean.
> >> > http://img844.imageshack.us/img844/8972/screenshotdx.png
> >> >
> >> > When I view S3_10_23 under Details view, it appears 8 regions.
> >> > http://img200.imageshack.us/img200/8070/screenshot1rr.png
> >> > Please help!!!
> >> > Thank you very much,
> >> >
> >> >
> >> > ------------------------------------------------------------------------------
> >> > This SF.net email is sponsored by
> >> >
> >> > Make an app they can't live without
> >> > Enter the BlackBerry Developer Challenge
> >> > http://p.sf.net/sfu/RIM-dev2dev
> >> > _______________________________________________
> >> > Gmod-gbrowse mailing list
> >> > [hidden email]
> >> > https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
> >> >
> >> >
> >>
> >>
> >>
> >> --
> >> ------------------------------------------------------------------------
> >> Scott Cain, Ph. D.                                   scott at scottcain
> >> dot net
> >> GMOD Coordinator (http://gmod.org/)                     216-392-3087
> >> Ontario Institute for Cancer Research
> >
> >
>
>
>
> --
> ------------------------------------------------------------------------
> Scott Cain, Ph. D.                                   scott at scottcain dot net
> GMOD Coordinator (http://gmod.org/)                     216-392-3087
> Ontario Institute for Cancer Research
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev _______________________________________________
> Gmod-gbrowse mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse





--
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research

------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
Reply | Threaded
Open this post in threaded view
|

Re: Search results are wrong. Please help.

wizi
Hi Scott, Lincoln, 
I got it. Thank you very much. I really appreciate it. 

On Tue, Aug 17, 2010 at 11:32 AM, Scott Cain <[hidden email]> wrote:
Hi Minh,

I wasn't advocating just swapping the Name and Alias, but having only one feature for the features that have the same boundaries, and putting the name values in the Alias, and putting the type in the Name.  Nevertheless, I don't think this is what you want to do if you still want to see 5 features when when there are 5 features that have the same boundaries.

I would suggest you follow the path that Lincoln suggested: assuming you aren't using the source (column 2) for anything else, you could put what is currently set as the name to the source, and then the features would be distinct when you do searches.

Scott



On Mon, Aug 16, 2010 at 2:07 PM, Minh Bui <[hidden email]> wrote:
Hi,

To Scott:  I try to swap Alias and Name but It still gives me the same results. 
To Chris:  Yes, its like BLAST but its much faster. I mapped >90,000 clusters to UCSC Human Genome for about 15-20 minutes. BLAST or BLAT may take days.  Here is the "Hits" table that summarize how many matches for each type (only on inactivated HIVs, there is Activated HIVs which is 4 times bigger than this) 
   
UNACTIVATED
Type Match
NL3_10_21 2909
NL3_10_23 482
NL3_11_11 965
S3_10_21 6962
S3_10_23 5245
S3_11_11 3219
W3_10_21 4643
W3_10_23 2360
W3_11_11 4148
Total 30933

Let me know if you can't see it. 


To Lincoln: when you say "different method or source", you mean the 2nd column in GFF format right? Could you be more specific? I am pretty new to Gbrowse2. 

Thank you all for your help. 

Minh Bui. 



On Mon, Aug 16, 2010 at 1:45 PM, Chris Fields <[hidden email]> wrote:
So this is akin to a seq analysis, with query/subject hits like BLAST, correct?  Have you looked at how matches and Target are defined in GFF3 (under "THE GAP ATTRIBUTE")?  You could also look into SAM/BAM and Bio-SamTools for this (which is pretty easy with GBrowse 2).

chris

On Aug 16, 2010, at 12:21 PM, Minh Bui wrote:

> Hi Scott,
> Here is how I come up with the GFF file:
> Step 1: I was given about ~100,000 clusters (HIV sequences were already trimmed off so I could map these clusters to Human genome. We want to know where HIVs integrated themselves into Human Genome)
> Step 2: I mapped these clusters to UCSC Human Genome using 454 GS Mapper.
> Step 3: The results were parsed to GFF using my Perl script.
>
> _There are 9 different types of HIV such as S3_10_23, W1_11_11.
> _My goal is to display all the locations of these types.
>
> In the note, I display the queries's sequences. And yes, they are different (they are all different on start-end positions). Some of them might have multiple hits on different chromosomes. Therefore, they might have the same "Notes" but different chromosome + start-end position .
>
>
>
> On Mon, Aug 16, 2010 at 12:53 PM, Scott Cain <[hidden email]> wrote:
> Hi Minh,
>
> So each of these features that share the same boundaries are different
> sequence variants?  I'm guessing that based on the sequence notes in
> the screenshot; the second and there sequences look nearly identical
> (I can see at least one snp though).
>
> I'm not sure how I would handle this.  On one hand, the search results
> are correct: it says the following 5 regions (not features) match, and
> if a user follows any one of those links, she will see all of the
> relevant features in the display.  I understand that it may mislead
> the user into thinking there are only 5 features when there are
> actually eight.
>
> To collapse the ones that share boundaries into one feature, you could
> swap the Name and Alias tags, where all of the features of a given
> type share the name of the type, and then the Alias would be populated
> with a list of all of the names of the individual GInumber-Type
> entries.  This might be less than ideal if you want to capture full
> sequence for each of the variants though.
>
> Scott
>
>
> On Mon, Aug 16, 2010 at 12:19 PM, Minh Bui <[hidden email]> wrote:
> > Hi Scott,
> > Could you explain more on collapsing into one feature?
> > Basically, what I want to know is the location of HIV clusters based on
> > type.
> > The clusters' names follow this format: GInumber-Type (Example: GI######,
> > Type: S3_10_23)
> > I want to be able to search all the location of a HIV-type.
> > How should I configure Gbrowse2 to do that?
> > Sorry about my english.
> > Let me know if you need more info.
> > Minh Bui.
> >
> > On Mon, Aug 16, 2010 at 11:44 AM, Scott Cain <[hidden email]> wrote:
> >>
> >> Hi Minh,
> >>
> >> I think what is happening here is that, when features of the same type
> >> share common boundaries, to the search tool, they look like the same
> >> result, particularly if the have the same aliases.  In fact, it seems
> >> to me that if that is the case, maybe the should all collapse into one
> >> feature anyway.  Is there a reason they aren't?
> >>
> >> Scott
> >>
> >>
> >> On Mon, Aug 16, 2010 at 10:49 AM, Minh Bui <[hidden email]> wrote:
> >> > Hi everyone,
> >> > I am trying to search with Alias name but the results are always less
> >> > than
> >> > its supposed to be. The GFF is loaded to mySQL using seqfeature_load.
> >> > When I
> >> > search for S3_10_23, it returns 5 regions instead of 8. It also happened
> >> > with other type (Alias) as well.
> >> > Please look at the screenshot below you will understand what I mean.
> >> > http://img844.imageshack.us/img844/8972/screenshotdx.png
> >> >
> >> > When I view S3_10_23 under Details view, it appears 8 regions.
> >> > http://img200.imageshack.us/img200/8070/screenshot1rr.png
> >> > Please help!!!
> >> > Thank you very much,
> >> >
> >> >
> >> > ------------------------------------------------------------------------------
> >> > This SF.net email is sponsored by
> >> >
> >> > Make an app they can't live without
> >> > Enter the BlackBerry Developer Challenge
> >> > http://p.sf.net/sfu/RIM-dev2dev
> >> > _______________________________________________
> >> > Gmod-gbrowse mailing list
> >> > [hidden email]
> >> > https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse
> >> >
> >> >
> >>
> >>
> >>
> >> --
> >> ------------------------------------------------------------------------
> >> Scott Cain, Ph. D.                                   scott at scottcain
> >> dot net
> >> GMOD Coordinator (http://gmod.org/)                     216-392-3087
> >> Ontario Institute for Cancer Research
> >
> >
>
>
>
> --
> ------------------------------------------------------------------------
> Scott Cain, Ph. D.                                   scott at scottcain dot net
> GMOD Coordinator (http://gmod.org/)                     216-392-3087
> Ontario Institute for Cancer Research
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev _______________________________________________
> Gmod-gbrowse mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse





--
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research


------------------------------------------------------------------------------
This SF.net email is sponsored by

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
Gmod-gbrowse mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gmod-gbrowse