On 22 May 2012 15:18, Gordon Gremme <gremme@gmail.com> wrote:
That's true, although the attribute Dbxref should be used for that purpose.

b) must have no meaning outside the file

I think the number of people who do not adhere to this is significant (including myself and VectorBase GFF3 files). Explicit tracking of names/IDs often takes place within separate fields (e.g. Name, Synonym, Locus) and not within the Dbxref field for many projects (see the FlyBase example). In fact I struggle to think of examples where the Dbxref contains self-referential links to the project from which the data was generated as opposed to third party databases.

cheers
Dan
 




On Tue, May 22, 2012 at 4:04 PM, Cook, Malcolm <MEC@stowers.org> wrote:
> FWIW, despite the documentation suggesting that the ID is only used to tie
> together features in transit within a GFF file, in my experience, in
> practice, the ID is additionally otherwise meaningful.
>
> For instance, Flyable uses their flybase identifier (FBgn#) as the ID of
> GFF lines for genes, i.e.:
>
>
>
>
>
> 2LHet   FlyBase gene    19633   32995   .       -       .       ID=FBgn0039958;Name=CG12567;Alias=FBan
> 0012567;...
>
> Not sure if this effects whatever decisions you're poised to makeŠ.
>
> Cheers,
>
> Malcolm
>
>
>
> On 5/22/12 8:39 AM, "Gordon Gremme" <gremme@gmail.com> wrote:
>
>>Hi Scott,
>>
>>> I agree with everything you wrote except this:
>>>
>>>> - IDs of features which are not multi-line features and don't have
>>>>  children can be dropped. That means that they can be processed
>>>>  instantaneously without the need of a terminator line (###).
>>>
>>> Here's the problem: how do you know if a given GFF line with an ID is
>>> not a multiline or has a child?  There is no way to know for sure, so
>>> you still need to keep track of the IDs until a terminator or EOF.
>>
>>Sorry if that was unclear. You are right, in the first parse one has
>>to wait until a terminator or EOF is encountered to make sure the
>>given line is not a multi-line feature or has a child. If that is the
>>case, the feature can be shown without the ID. And then in the next
>>parse it could be processed instantaneously, because it doesn't have
>>an ID anymore and therefore can't be a multi-line feature or have
>>children. That is, this advantage comes only into play after
>>superfluous ID attributes have been dropped.
>>
>>I hope that clears it up,
>>
>>Gordon
>>
>>--------------------------------------------------------------------------
>>----
>>Live Security Virtual Conference
>>Exclusive live event will cover all the ways today's security and
>>threat landscape has changed and how IT managers can respond. Discussions
>>will include endpoint security, mobile security and the latest in malware
>>threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
>>_______________________________________________
>>SOng-devel mailing list
>>SOng-devel@lists.sourceforge.net
>>https://lists.sourceforge.net/lists/listinfo/song-devel
>
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> _______________________________________________
> SOng-devel mailing list
> SOng-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/song-devel

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
SOng-devel mailing list
SOng-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/song-devel



--
Ensembl Genomes | VectorBase | i5K insect genome initiative