From GO Wiki
Jump to: navigation, search

You are now talking on #AmiGO

<sjcarbon> Sorry that I'm late!
<livstone> Welcome!
<rama> no problem
<sidd> Hi seth
<sjcarbon> Hello!
Right to business...
Has everybody had a chance to take a look at the mock-ups?
<livstone> Yes.
<rama> yes
<sjcarbon> So, what do y'all think? Just shout it out.
Any combinations?
Something more that you want to see?
<livstone> Honestly, I like the original the best, but I'm still kinda fuzzy as to what the PURPOSE of the page is.
<pascale> i think that covers all suggestions
<rama> Thanks for mocking those versions. 
<sjcarbon> No problem.
I think the purpose might be better covered over email, 
unless anybody wants to take a stab at it right now?
<pascale> should we have people vote on this? 
how are we going to pick?
<sjcarbon> One quick thing first:
Pascale, you mentioned (in email) that you like "heat map clumped", but want to see the GPs.
But since the whole block is filled-in, we wouldn't really be able to do that--it would look more like "heat map".
<pascale> I thought all the gene names could be listed
<livstone> I agree that all gene names should be listed.
<sjcarbon> But the different genes could have different evidence, and thus different backgrounds.
<rama> On the heat map clamp page, I would prefer to see the gene names instead of just Human, fly etc
<sjcarbon> But having a mixed bunch be all green might be misleading.
<rama> may be make EXP as red and ISS as blue
<livstone> I think each gene should get its own evidence code.
<sjcarbon> Actually, how about everybody just shouts out their favorite, and then we can talk details.
<pascale> sounds good 
<rama> I like the original version
(Sorry pascale!)
<sjcarbon> original: 1
<rama> very few of us to vote here
<pascale> right 
<sjcarbon> true, but it might help the discussion
<livstone> Again, I like the original the best.  (each gene listed individually, with its own evidence code; no "OTHER" tag)
<pascale> np!!
<rama> Did more people comment about these pages on the ref.genome mailing list?
<pascale> well we were not to list all evidence codes
<rama> No body else is voting??
<sjcarbon> i love them all equally.
<pascale> :)
<livstone> They're like children :-)
<rama> Hah! 
<sjcarbon> maybe we should wait for a little comment from the list, but it looks like the original and "clumped" are the early leads
clumped being the most like Mary's pages.
<rama> I like the Heatmap-clumped, but few other things need to be fixed on that I feel. 
it is not that intuitive to a joe user
<livstone> I have 2 concerns with "clumped"...
<sjcarbon> how about "heat map"?
livstone: go on
<pascale> i am not sure any of this is intuitive; we need joe testers
<rama> can we go with the original for now and get more comments
<livstone> 1) The requested changes make it look more like the original, and
<rama> yes, none of these is that intuitive..but the whole thing is very hard to conceptualize 
<livstone> 2) It would be hard for people who are color-bloind to read the heat map.
<rama> and who is the audience for this
<sjcarbon> livstone: another reason we went with the labels initially.
(actually, not initially, but we did eventually do it)
<livstone> Rama brings up a good point.  Maybe we should talk a bit about the purpose of this page for a minute?
<rama> I think we have spent way too much time on this
this page should serve as an introduction/ quick view on what are the orthologs
and i think the original page serves that purpose
there is always a concept called 'enhancement'
if we want to change displays
<livstone> Is it for us to keep track of the curation status of individual genes in families, or is it for the scientific public to see what cool stuff we have?
<rama> this page is mostly for the public (that is my understanding)..
curators can of course use it
<sjcarbon> the RG that sidd is working on is more for internal stuff
this webpage will be mostly for external consumption
<livstone> OK, cool.  Then I definitely think it's cool in the original form, and I'm OK with EXP trumping ISS and ISS trumping TAS/NAS/none.
Seth, I sent some suggested wording changes for the key last night.  Have you gotten them?
<sjcarbon> ...
yes, thank you!
<livstone> Great.  Thanks.  Do they need discussion, or should we not bother?
<sjcarbon> by the way--I have a response from varsha...
prefers "heat map"--basically the most information-rich display.
<livstone> So, the difference between heat map and the original is (1) colors for the tags, and (2) the use of the "OTHER" tag, right?

--> stacia (stacia@adsl-99-139-253-68.dsl.skt2ca.sbcglobal.net) has joined #AmiGO

<sjcarbon> the difference between current and verbose is the use of "OTHER"
<stacia> m
<sjcarbon> (or whatever we would want to call it)
the difference between verbose and "heat map" is the use of color and differentiating evidence codes
oops--skip the differentiating part.

<-- stacia has quit (Quit)

<sjcarbon> (too much for poor stacia I'm afraid)
<livstone> OK.  Well, I think that we should just drop the "OTHER" tag entirely. Show the gene name, of course, but leave it unmentioned that there is another type of evidence code.
<pascale> and how would we distinguish that from no annotations? 
<livstone> We wouldn't.
<sjcarbon> it is a loss of information, but one could argue that it is not useful information in this context.
<livstone> Agreed.  What are the current possibilities for "OTHER", anyway?
<rama> other ev-codes like NAS, TAS?
<sjcarbon> TAS, oddly formed ISS...
...anything else that might be odd or have gone wrong.
<livstone> Yeah...I wouldn't want to publicize those too much.
<pascale> well another point is that the intro to that page sayd (or used to) that the genes were 'completely annotated'
so they should NOT have 'other' (TAS or oddly formed ISS)
so maybe it's okay to group those annotations with 'no annotations'
<sjcarbon> The text is an open book for all of you. If you send (to the list), I will put it on the page.
<livstone> I agree, Pascale, that they should be grouped with "no annotations," and I think that the way we designate that is not to show an evidence code.
<pascale> i would prefer saying something just so people dont see it as an omission (none?)
<livstone> Here is the text I sent Seth last night:
In each family, any homologs from a given species are shown.  Evidence codes, if present, indicate the folowing:
? EXP: The homolog has at least one annotation directly supported by experimental evidence.
? ISS: The homolog has at least one annotation supported by sequence similarity, but none supported by experimental evidence.
? (Optional) No evidence code: No EXP or ISS annotations.
<pascale> what about IC
<livstone> Good question.  What do you think>?
<sjcarbon> This display is to show experimental evidence, right?
<pascale> I am not sure everyone agrees on that
I dont know
I feel there is not enough people to make a decision
<sjcarbon> Can IC mean something that is not direct experimental evidence?
<pascale> well sure
<livstone> Well, last week, I suggested that we not even show the ISS's.
<pascale> or else it would be annotated with an exp code
<sjcarbon> At least with ISS, there should be a bridge to direct experimental evidence, right?
<pascale> well mike- that's why I like the heat map- it gives some idea of the quality of the data everywhere, without trying to overload
ISS yes
IC, not too sure
I know we have IC from ISS
<sjcarbon> Sounds like it's weaker.
<pascale> I think that's accepted according to the guidelines
<sjcarbon> We can always add it later.
<pascale> we would need another evidence code :)
seth-- what about question #1: are we having this page at all??
<livstone> Pascale, I'm good with a heat map, provided that (1) each gene gets its own evidence code, and (2) the map is very very legible.
<sjcarbon> pascale: I see no reason at this point to drop it. It will be a bit buried.
<pascale> why do you need each gene to have an evidence code? 
the table talks about homolog sets
<rama> I can go with heat map only if each gp name is listed in each cell and is made clickable
<pascale> right
<rama> (Pascale, are you agreeing with me?)
<livstone> Sometimes you get multiple gps from an organism with different evidence codes.  If each organism in a family gets 1 heat map color, then that information is lost,
<rama> This page is to not meant to be the be all page.
<sjcarbon> for that reason, I'm a little uncomfortable with "clumped"
<rama> It is to introduce the user to the whole concept
<livstone> Yeah, that's why I'm uncomfortable with "clumped," too.
<rama> we should be very careful about overloading
<pascale> (rama-- yes I agree about links to each gene in the set)
<rama> I am fine with the hm.html
but not the heat map clumped
Saying 'Other' is fine with me..

<-- sidd (sidd@reserved-dhcp-165-124-223-116.nubic.northwestern.edu) has left #AmiGO (Leaving) livstone has quit (Read error: Connection reset by peer)

<sjcarbon> the only difference between verbose and "heat map" are the colors.

--> Mike (livstone@vpn2-client-a37.Princeton.EDU) has joined #AmiGO

<sjcarbon> time is up.
<Mike> Hi.  I got disconnected.  did I miss anything?
<sjcarbon> not too much.
let's let it sit for a day and see if it gets a response.
<pascale> okay
<Mike> OK
<rama> Seth,
you should post it on the two mailing lists and put a time dead line for voting..
<pascale> maybe you sould email again and say 'please comment by tomorrow, we;ll implement the current version if there are no further objections'
<rama> I would also remove the heat map clumped
from the list. 
<sjcarbon> and poor old minimal?
<rama> too many choices
<Mike> Yeah, no minimal
<rama> ask people to vote between original, verbose and heatmap
<sjcarbon> no problem.
if there is no other business, shall we wrap-up?
<rama> thanks. and sorry this is taking so much time

<-- cjm has quit (Quit)

<rama> yes, wrap up
<sjcarbon> bye!
<Mike> Thanks!
<pascale> bye
<Mike> Bye!

<-- pascale has quit (Quit)

Mike (livstone@vpn2-client-a37.Princeton.EDU) has left #AmiGO
rama (rama@c-24-7-118-198.hsd1.ca.comcast.net) has left #AmiGO