search

qiime / qiime_web_app

QIIME, EMP and American Gut Web Application
www.microbio.me/qiime
9 stars 10 forks source link

Add ability to color plots based on taxonomy data #116

Open douginator2000 opened 11 years ago

douginator2000 commented 11 years ago

PCoA 2D/3D, others? Bi-plots?

douginator2000 commented 11 years ago

Probably will just show up as additional fields in the mapping file for varying levels of taxonomy data.

Question: There will of course be many taxa associated with a particular sample. If you color the sample's "dot", do you pick the most abundant taxa to color it by?

rob-knight commented 11 years ago

Yes, all of these. User must specify taxon/taxa to color by.

On Jan 28, 2013, at 2:14 PM, "Doug Wendel" notifications@github.com<mailto:notifications@github.com> wrote:

PCoA 2D/3D, others? Bi-plots?

— Reply to this email directly or view it on GitHubhttps://github.com/qiime/qiime_web_app/issues/116.

rob-knight commented 11 years ago

Better to let user specify, need not be most abundant.

On Jan 28, 2013, at 2:17 PM, "Doug Wendel" notifications@github.com<mailto:notifications@github.com> wrote:

Probably will just show up as additional fields in the mapping file for varying levels of taxonomy data.

Question: There will of course be many taxa associated with a particular sample. If you color the sample's "dot", do you pick the most abundant taxa to color it by?

— Reply to this email directly or view it on GitHubhttps://github.com/qiime/qiime_web_app/issues/116#issuecomment-12805287.

douginator2000 commented 11 years ago

Chatted with Cathy about this: Need a way to select which taxa and and at what level prior to performing analyses. Don't really know what to select until looking at output of summarize_taxa.py. Think about how to work this data into a good UI.

Also confirmed that the coloring is generally a color gradient based on abundance of selected taxa in each sample. Also could be interesting to color by most abundant taxa in some way but this is a secondary item.

wasade commented 11 years ago

what about 2 pull downs, first to select level, second to select name observed at that level?

On Mon, Jan 28, 2013 at 3:25 PM, Doug Wendel notifications@github.comwrote:

Chatted with Cathy about this: Need a way to select which taxa and and at what level prior to performing analyses. Don't really know what to select until looking at output of summarize_taxa.py. Think about how to work this data into a good UI.

— Reply to this email directly or view it on GitHubhttps://github.com/qiime/qiime_web_app/issues/116#issuecomment-12808980.

douginator2000 commented 11 years ago

Sounds like a reasonable end-game though given that the plot parameters are currently derived directly from the Qiime script params dict, it might make sense to add this ability directly to Qiime (e.g. add an additional optional input file and the summarization level/taxa), or we may need to think about exactly how to add this coloring option as an add-on that the website provides.

On Mon, Jan 28, 2013 at 3:29 PM, Daniel McDonald notifications@github.comwrote:

what about 2 pull downs, first to select level, second to select name observed at that level?

On Mon, Jan 28, 2013 at 3:25 PM, Doug Wendel notifications@github.comwrote:

Chatted with Cathy about this: Need a way to select which taxa and and at what level prior to performing analyses. Don't really know what to select until looking at output of summarize_taxa.py. Think about how to work this data into a good UI.

— Reply to this email directly or view it on GitHub< https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12808980>.

— Reply to this email directly or view it on GitHubhttps://github.com/qiime/qiime_web_app/issues/116#issuecomment-12809134.

Doug Wendel Professional Research Associate University of Colorado at Boulder Department of Chemistry and Biochemistry, Knight Lab 303-641-5442 wendel@colorado.edu

wasade commented 11 years ago

But taxon data is in biom table already

On Jan 28, 2013, at 15:41, Doug Wendel notifications@github.com wrote:

Sounds like a reasonable end-game though given that the plot parameters are currently derived directly from the Qiime script params dict, it might make sense to add this ability directly to Qiime (e.g. add an additional optional input file and the summarization level/taxa), or we may need to think about exactly how to add this coloring option as an add-on that the website provides.

On Mon, Jan 28, 2013 at 3:29 PM, Daniel McDonald notifications@github.comwrote:

what about 2 pull downs, first to select level, second to select name observed at that level?

On Mon, Jan 28, 2013 at 3:25 PM, Doug Wendel notifications@github.comwrote:

Chatted with Cathy about this: Need a way to select which taxa and and at what level prior to performing analyses. Don't really know what to select until looking at output of summarize_taxa.py. Think about how to work this data into a good UI.

— Reply to this email directly or view it on GitHub< https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12808980>.

— Reply to this email directly or view it on GitHub< https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12809134>.

Doug Wendel Professional Research Associate University of Colorado at Boulder Department of Chemistry and Biochemistry, Knight Lab 303-641-5442 wendel@colorado.edu

— Reply to this email directly or view it on GitHubhttps://github.com/qiime/qiime_web_app/issues/116#issuecomment-12809736.

douginator2000 commented 11 years ago

Correct though the challenge is giving the user enough information to select the appropriate things to color by prior to generating the charts/plots. E.g. Cathy says she needs to look at the output of summarize_taxa before she can make good choices about what taxa to color by.

Another thought: perhaps we could generate a number of plots/charts at various taxa levels and add some number of new columns to the mapping file for the most abundant taxa.

On Mon, Jan 28, 2013 at 3:48 PM, Daniel McDonald notifications@github.comwrote:

But taxon data is in biom table already

On Jan 28, 2013, at 15:41, Doug Wendel notifications@github.com wrote:

Sounds like a reasonable end-game though given that the plot parameters are currently derived directly from the Qiime script params dict, it might make sense to add this ability directly to Qiime (e.g. add an additional optional input file and the summarization level/taxa), or we may need to think about exactly how to add this coloring option as an add-on that the website provides.

On Mon, Jan 28, 2013 at 3:29 PM, Daniel McDonald notifications@github.comwrote:

what about 2 pull downs, first to select level, second to select name observed at that level?

On Mon, Jan 28, 2013 at 3:25 PM, Doug Wendel notifications@github.comwrote:

Chatted with Cathy about this: Need a way to select which taxa and and at what level prior to performing analyses. Don't really know what to select until looking at output of summarize_taxa.py. Think about how to work this data into a good UI.

— Reply to this email directly or view it on GitHub< https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12808980>.

— Reply to this email directly or view it on GitHub< https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12809134>.

Doug Wendel Professional Research Associate University of Colorado at Boulder Department of Chemistry and Biochemistry, Knight Lab 303-641-5442 wendel@colorado.edu

— Reply to this email directly or view it on GitHub< https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12809736>.

— Reply to this email directly or view it on GitHubhttps://github.com/qiime/qiime_web_app/issues/116#issuecomment-12810014.

Doug Wendel Professional Research Associate University of Colorado at Boulder Department of Chemistry and Biochemistry, Knight Lab 303-641-5442 wendel@colorado.edu

wasade commented 11 years ago

Correct though the challenge is giving the user enough information to

select the appropriate things to color by prior to generating the charts/plots. E.g. Cathy says she needs to look at the output of summarize_taxa before she can make good choices about what taxa to color by.

they cant look at a taxa summary plot first and then continue the analysis? or redo betadiv?

Another thought: perhaps we could generate a number of plots/charts at various taxa levels and add some number of new columns to the mapping file for the most abundant taxa.

agree with rob that user selection is ideal. most abundant may not be the most interesting

On Mon, Jan 28, 2013 at 3:48 PM, Daniel McDonald notifications@github.comwrote:

But taxon data is in biom table already

On Jan 28, 2013, at 15:41, Doug Wendel notifications@github.com wrote:

Sounds like a reasonable end-game though given that the plot parameters are currently derived directly from the Qiime script params dict, it might make sense to add this ability directly to Qiime (e.g. add an additional optional input file and the summarization level/taxa), or we may need to think about exactly how to add this coloring option as an add-on that the website provides.

On Mon, Jan 28, 2013 at 3:29 PM, Daniel McDonald notifications@github.comwrote:

what about 2 pull downs, first to select level, second to select name observed at that level?

On Mon, Jan 28, 2013 at 3:25 PM, Doug Wendel notifications@github.comwrote:

Chatted with Cathy about this: Need a way to select which taxa and and at what level prior to performing analyses. Don't really know what to select until looking at output of summarize_taxa.py. Think about how to work this data into a good UI.

— Reply to this email directly or view it on GitHub<

https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12808980>.

— Reply to this email directly or view it on GitHub< https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12809134>.

Doug Wendel Professional Research Associate University of Colorado at Boulder Department of Chemistry and Biochemistry, Knight Lab 303-641-5442 wendel@colorado.edu

— Reply to this email directly or view it on GitHub< https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12809736>.

— Reply to this email directly or view it on GitHub< https://github.com/qiime/qiime_web_app/issues/116#issuecomment-12810014>.

Doug Wendel Professional Research Associate University of Colorado at Boulder Department of Chemistry and Biochemistry, Knight Lab 303-641-5442 wendel@colorado.edu

—

Reply to this email directly or view it on GitHubhttps://github.com/qiime/qiime_web_app/issues/116#issuecomment-12810361.