These are chat archives for BD2K-DDI/specifications

18th
Mar 2015
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 09:30
Did you see this message
Mingze Bai
@baimingze
Mar 18 2015 09:41
yes
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 09:49
I was reviewing yesterday the mockups and the main think we should decide is if we will show (very begining) for evey search the list of resources one by one of the index. LEt says we have the https://raw.githubusercontent.com/PSI-PROXI/datadiscoveryindex/master/docs/DDI%20Browser%20Home%20page.png a search for cancer we will show the first hit for every resource
the more complicate approach could be like this:
but it over complica the search results page.. Also henning think it could be good to have tabs in the web like:
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:00
@baimingze about the filtering The first one should be the categories like @baimingze propouse. After that I prefer the Repositories because all the datasets will contain the repository where they are hosted.
Mingze Bai
@baimingze
Mar 18 2015 10:01
I agree
here we show the FIRST and SECOND entry of each omic
?
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:02
The scond one should be Species like we propouse because it really nice and also most of the providers should contains the specie, actually @baimingze we should make it mandatory in the BD2K schema and for Metabologiths, PRIDE and EGA
Mingze Bai
@baimingze
Mar 18 2015 10:07
should we change the page to the tabs?
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:08
no .. I like this aproach and we can keep this one
Mingze Bai
@baimingze
Mar 18 2015 10:09
I am not sure about the represent page after the clicking of the : < > 1-1309 results
we need a new window to show all the 1309
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:10
you mean for evry resource a pagination
Mingze Bai
@baimingze
Mar 18 2015 10:10
yes, we need pagination for every resource
and it is as the same as we clicking on the color checkbox on the topleft
AND the "sort by :" function is not needed in this page, but in each omic's page
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:15
ok I’m doing another refine in the page because we need to define waht is refine and what is Show results for like amazon: http://www.amazon.co.uk/s/ref=nb_sb_ss_sc_1_5?url=search-alias%3Daps&field-keywords=wheel&sprefix=whell%2Caps%2C155
We should use a Result for (for the categories), species, etc. and the filte for those filters we don’t know if they are present in all the datasets
Mingze Bai
@baimingze
Mar 18 2015 10:31
sorry, I can't follow you
I just found you changed the check boxes' positions
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:32
The "Show Results for” will look like categories like Repositories Species and topics
The Refine by will act like a filter independently of the category for example Instruments etc
If you are agree with this first mockup @baimingze we can send it to other developers and close the issue
Mingze Bai
@baimingze
Mar 18 2015 10:35
"Show Results for" vs "Refine by", I got it
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:35
At least for the Search Page
Mingze Bai
@baimingze
Mar 18 2015 10:36
The last one , we don't need the pagination, am I right?
at the bottom
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:38
no it would show one for every resource. Im the future we can implement the ranking
Mingze Bai
@baimingze
Mar 18 2015 10:39
ok
I agree
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:40
the user can always go inf they are interested to the resources or field by cliking in the REsult for
Mingze Bai
@baimingze
Mar 18 2015 10:43
In my first thought, all of them are in "AND" operation, Just like the Metabolights
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:43
If you are agree I will create an issue and close it with the final mockup
yes sure
Mingze Bai
@baimingze
Mar 18 2015 10:44
now I am not under stand the difference between "show results for" and "refine by"
is that still all "AND"?
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:45
but for example imagine you have cancer search and then you are interested only in the metaboligths you basically click on repository metaboligths and you will only have this view
Mingze Bai
@baimingze
Mar 18 2015 10:45
yes, then I can still check on the Species, Instrument
how different between "Sepecies" and "Instrument"?
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:46
Show results for is more a category basic filter if you are in Proteomics you will not be in Metaboligths
if you are human are in peptideAtlas you shouldn’t be in PRIDE
Mingze Bai
@baimingze
Mar 18 2015 10:47
I got
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:47
perhaps is better to also move Species to Filter by
Mingze Bai
@baimingze
Mar 18 2015 10:47
Then "Species" should be in "Refine by", because some
yes
that's what I mean
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:48
ok I will do it
Mingze Bai
@baimingze
Mar 18 2015 10:48
OK.
no questions now
any thing else can I do?
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:50
I will commit the code
Mingze Bai
@baimingze
Mar 18 2015 10:51
OK, then I learn some JQuery
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:51
already pdated
We can talk with Rui about the technology
Mingze Bai
@baimingze
Mar 18 2015 10:52
Yean, I have to learn the basic concepts
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 10:52
great
@pconesa welcome to the discussion
Pablo Conesa
@pconesa
Mar 18 2015 11:02
Hi, I just read about the species....
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:03
A quick update for @pconesa .. after some discussions with external services and because in the first step only Metaboligths, PRIDE and EGA will be involve in the project, we decided to use the EBE_EYE file we are generating for the EBI search. It means that all the work related with parsing, index and .. will be in the EBI Service Team. Noew we are working in the mockups for the web. WE included a filter and group version in the web #1 .. It would be great if you can help us to define some of them for metaboligths, for example are you using Organism part.. are you @pconesa exporting this information to EBE search.
Pablo Conesa
@pconesa
Mar 18 2015 11:03
It's very triky to have CV on species, at least for us and can't be mandatory
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:04
@pconesa but if you have this information at some point can you expor it and use the proper filed for that
Pablo Conesa
@pconesa
Mar 18 2015 11:07
We have organism Part and Organism but in some cases makes sense to have it both empty. We have some studies done in "sea water" and environmental samples (soil, water...) are accepted.
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:09
doesn't matter @pconesa if the information is not provider when the user click the filter the experiments that do not contains this information will be removed from the search is a matter of high-quality annotations …
but if some of them contains these annotations it would be great to have the clasification
Pablo Conesa
@pconesa
Mar 18 2015 11:12
That would work.
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:12
We are following the aproach of amazon if the information is present we boost the result if not we remove it
@pconesa @baimingze another thing which information are you exporting to EBE_EYE file
Pablo Conesa
@pconesa
Mar 18 2015 11:16
EBI export fields:
id
last_modification_date
name
organism
platform
study_design
study_factor
study_status
submitter
technology_type
Mingze Bai
@baimingze
Mar 18 2015 11:18
@pconesa some of these field are for MTBC, NOT for MTBS, am I right?
Pablo Conesa
@pconesa
Mar 18 2015 11:19
Yes I removed non relevant ones
@baimingze , you are right
Mingze Bai
@baimingze
Mar 18 2015 11:20
organism is like Species, can we choose a term to represent of them?
and the "Instrument" vs "Platform"
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:22
@baimingze @pconesa the idea is to standarize the interface even when we don’t modify for now the XML files
but in the proposal default schema for BD2K we should propose the one we chose
Mingze Bai
@baimingze
Mar 18 2015 11:24
ok
Pablo Conesa
@pconesa
Mar 18 2015 11:24
yes organism ~= Species (in some of our cases it's not an organism, but it's shown under Organism...maybe something more generic will be bertter like "Source"....any suggestion?
instrument is platform...yes @baimingze
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:25
If we use ontolgy definion species are a global term and orgnisms are inside http://en.wikipedia.org/wiki/Species
then we can keep species
in the interface we can use Species & Organisms
in the file we should use the global term species
what do you think
Pablo Conesa
@pconesa
Mar 18 2015 11:26
@ypriverol what about "marine water"...this is not an species
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:26
but also is not and organism
Pablo Conesa
@pconesa
Mar 18 2015 11:27
yep, but if a user is looking for studies done in soil
I'm just trying to say that our Sample concept (is broader than species)
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:28
@pconesa if I’m understand you are putting information about the source in organism doesn’t matter if is an organism or not
?
Pablo Conesa
@pconesa
Mar 18 2015 11:29
So far, yes...it's wrong we know that and want to fix it
Let's say it's not that controlled, ;-)
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:30
do you have a way to detect fo those cases that are Organisms and species define them like it and for those let says “general source” like others?
Pablo Conesa
@pconesa
Mar 18 2015 11:33
yes we do
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:33
it would be like splt the information to have a better annotation
I also see that ORGANIMS part is like a cellular component @pconesa is that a CVTerm
Pablo Conesa
@pconesa
Mar 18 2015 11:38
It could be Liver, Leaf, so far is more an organ/anatomy than a cellular component but it could be in case someone does an Metabolomics study on mitochondrial fluid (if possible!)
@ypriverol we have alway the free text option...but in many cases it's a CV term
Mingze Bai
@baimingze
Mar 18 2015 11:40
Then the filed "Tissue" in PRIDE has relationship with "Organism" too?
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:41
@baimingze Tissue is one thing and organism is another one organism can be rat or human and the tissue can be still liver tissue for both
@pconesa how we can standarize in some way the Organism Part
Mingze Bai
@baimingze
Mar 18 2015 11:43
according to @pconesa 's words, the content in filed "Organism" could be "Liver", which is a "Tissue".
Pablo Conesa
@pconesa
Mar 18 2015 11:45
no, Organism can't be liver
Pablo Conesa
@pconesa
Mar 18 2015 11:46
we have Organism Part (not exported to ebi). It's a different field. We have "Organism" and we have "Organism Part"
Mingze Bai
@baimingze
Mar 18 2015 11:46
oops :smile:
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 11:47
@pconesa instead of Organism Part we can have someting like "Organism part” and for those that you already know they are Tissue or Cellular component has like a reference
Yasset Perez-Riverol
@ypriverol
Mar 18 2015 15:26
@baimingze I was talking with @noedelta and one of the key issues here will be to do facet for multi lucene indexes
can you ask to Luciano
Silvano
Mingze Bai
@baimingze
Mar 18 2015 15:43
ok.