I think it works both ways, as long as you don't store duplicate experiment-phenotype values .<br><br>I think an nd_experiment should be a phenotyping/genotyping event, with all measurements sharing the same properties (like date, location, person, etc.) <br>
<br>Of course if someone does not agree with this definition, there is no problem storing one experiment id for each single measurement. The 'experiment' notion in the natural div. module is just a way for linking stocks with phenotypes/genotypes, while allowing to re-use the same stock in multiple projects or field plots or whatever. <br>
<br>I guess the VectorBase people could chip in about making this work in the opposite direction ( an experiment yields new stocks) <br><br>-Naama<br><br><br><div class="gmail_quote">On Fri, Oct 22, 2010 at 11:58 AM, Sook Jung <span dir="ltr"><<a href="mailto:sookjc@gmail.com">sookjc@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">Hi All,<br>
I think I raised this issue many times but it somehow gets forgotten..<br>
We decided long time ago that nd_experiment to phenotype was many to<br>
one, not one to one, since many experiment can be linked to one<br>
phenotype. For example, many experiments with different samples can<br>
have sugar content of '3', so we didn't want to make multiple raws in<br>
phenotype table for each experiment. So 1-1 constraint on<br>
(nd_experiment_id, phenotype_id) would't work..<br>
<br>
Then some other people suggested linking one experiment to many<br>
phenotypes which is different isssue.<br>
<br>
Thanks<br>
Sook<br>
<div><div></div><div class="h5"><br>
On Fri, Oct 22, 2010 at 8:21 AM, Naama Menda <<a href="mailto:nm249@cornell.edu">nm249@cornell.edu</a>> wrote:<br>
> Good point.<br>
> I dropped the 1-1 relationship and added a constraint on (nd_experiment_id,<br>
> phenotype_id /genotype_id )<br>
><br>
> It's in svn, so update your schema if you're using it .<br>
><br>
> -Naama<br>
><br>
><br>
> Naama Menda<br>
> Boyce Thompson Institute for Plant Research<br>
> Tower Rd<br>
> Ithaca NY 14853<br>
> USA<br>
><br>
> (607) 254 3569<br>
> Sol Genomics Network<br>
> <a href="http://solgenomics.net/" target="_blank">http://solgenomics.net/</a><br>
> <a href="mailto:nm249@cornell.edu">nm249@cornell.edu</a><br>
><br>
><br>
> On Fri, Oct 22, 2010 at 9:27 AM, Hilmar Lapp <<a href="mailto:hlapp@nescent.org">hlapp@nescent.org</a>> wrote:<br>
>><br>
>> It looks to me that after relaxing this uniqueness constraint the table<br>
>> doesn't have a natural primary key anymore. Every table should have a<br>
>> natural primary key enforced by a UNIQUE constraint. Maybe the constraint<br>
>> should be on (nd_experiment_id,phenotype_id)?<br>
>> -hilmar<br>
>> On Oct 22, 2010, at 7:46 AM, Naama Menda wrote:<br>
>><br>
>> hi Bob,<br>
>> you could store the 2 measurements with 2 different experiment ids, or use<br>
>> the same experiment for both.<br>
>> Someone at some point asked for the 1-1 constraint, and I can't remember<br>
>> why.<br>
>> I think it's best to have users enforce this if needed on the software<br>
>> level.<br>
>><br>
>> Any other thoughts?<br>
>><br>
>> -Naama<br>
>><br>
>><br>
>><br>
>> On Fri, Oct 22, 2010 at 5:53 AM, Bob MacCallum<br>
>> <<a href="mailto:r.maccallum@imperial.ac.uk">r.maccallum@imperial.ac.uk</a>> wrote:<br>
>>><br>
>>> Wasn't it just that if you are measuring two phenotypes, say leaf size<br>
>>> and fruit weight, then you'd use two nd_experiments (each linked to the same<br>
>>> stock) because you need two nd_protocols to describe the two assays?<br>
>>><br>
>>> We have no objections to relaxing the constraint though - I'm sure there<br>
>>> are situations where 1 to many is needed.<br>
>>><br>
>>><br>
>>> On Thu, Oct 21, 2010 at 9:00 PM, Naama Menda <<a href="mailto:nm249@cornell.edu">nm249@cornell.edu</a>> wrote:<br>
>>>><br>
>>>> hi,<br>
>>>><br>
>>>> does anyone know why there was a request to have the experiment_id<br>
>>>> unique constraint on nd_experiment_phenotype ad genotype tables?<br>
>>>><br>
>>>> There was a request recently to drop this 1-1 relationship, and I can't<br>
>>>> figure out why it was requested in the first place .<br>
>>>><br>
>>>> thanks!<br>
>>>> -Naama<br>
>>>><br>
>>>><br>
>>>> Naama Menda<br>
>>>> Boyce Thompson Institute for Plant Research<br>
>>>> Tower Rd<br>
>>>> Ithaca NY 14853<br>
>>>> USA<br>
>>>><br>
>>>> (607) 254 3569<br>
>>>> Sol Genomics Network<br>
>>>> <a href="http://solgenomics.net/" target="_blank">http://solgenomics.net/</a><br>
>>>> <a href="mailto:nm249@cornell.edu">nm249@cornell.edu</a><br>
>>><br>
>><br>
>><br>
>> --<br>
>> ===========================================================<br>
>> : Hilmar Lapp -:- Durham, NC -:- <a href="http://informatics.nescent.org" target="_blank">informatics.nescent.org</a> :<br>
>> ===========================================================<br>
>><br>
>><br>
><br>
><br>
<br>
<br>
<br>
</div></div><font color="#888888">--<br>
Sook Jung, PhD<br>
Assistant Research Professor of Bioinformatics<br>
Dept of Horticulture and Landscape Architecture<br>
Washington State University<br>
45 Johnson Hall, Pullman, WA 99164-6414<br>
<a href="mailto:Email%3Asook@bioinfo.wsu.edu">Email:sook@bioinfo.wsu.edu</a><br>
</font></blockquote></div><br>