New Google Group for Spark
49 views
Skip to first unread message
Cydney Nielsen
Jun 8, 2012, 4:10:31 PM6/8/12
to Spark - Genome-Scale Clustering and Visualization
Welcome to the Spark Google Group! This is an initial test post.
seqfast
Nov 10, 2012, 4:20:40 PM11/10/12
to spark...@googlegroups.com
Hi Cydney,
I'm getting ready to use Spark, it seems it can do a lot of what I've been doing manually in a much nicer fashion. It's a great site and the video tutorials are wonderful to have.
Quick question, more of a feature 'do you have it' or request. For each of the clusters that are generated for the various tracks, is it possible to export the various tracks' signal as a line plot, like the ones we are used to see in many papers (e.g. 5kb up down tss)? Since the heatmap/track map for the cluster represents the average (?) signal, it seems easy enough to plot it. As well, some kind option for average/median plotting, or something with percentile/confidence intervals would be great.
While i'm asking away, something I end up doing a lot that would nice to have more automated is to take those clusters for those regions of interest (peaks, tss, enhancers, etc) and rank order them by the occupancy of one of the tracks - then display them as a heatmap with each tss/peak having it's own line. Usually we generate per-gene tss plots, or gene body plots, and then use cluster (or order by tss occupancy or something) and view with Treeview. This works well enough, but having the option to have these all in one place would be fabulous.
Either way, I'm looking forward to getting familiar with Spark, and thanks for the development of a promising tool!
-sf
I'm getting ready to use Spark, it seems it can do a lot of what I've been doing manually in a much nicer fashion. It's a great site and the video tutorials are wonderful to have.
Quick question, more of a feature 'do you have it' or request. For each of the clusters that are generated for the various tracks, is it possible to export the various tracks' signal as a line plot, like the ones we are used to see in many papers (e.g. 5kb up down tss)? Since the heatmap/track map for the cluster represents the average (?) signal, it seems easy enough to plot it. As well, some kind option for average/median plotting, or something with percentile/confidence intervals would be great.
While i'm asking away, something I end up doing a lot that would nice to have more automated is to take those clusters for those regions of interest (peaks, tss, enhancers, etc) and rank order them by the occupancy of one of the tracks - then display them as a heatmap with each tss/peak having it's own line. Usually we generate per-gene tss plots, or gene body plots, and then use cluster (or order by tss occupancy or something) and view with Treeview. This works well enough, but having the option to have these all in one place would be fabulous.
Either way, I'm looking forward to getting familiar with Spark, and thanks for the development of a promising tool!
-sf
Cydney Nielsen
Nov 13, 2012, 4:32:16 PM11/13/12
to spark...@googlegroups.com
Hi sf,
Thanks for your feedback and interest in Spark!
Both of your feature requests - (1) summary line plots for each cluster, and (2) sortable heatmaps displaying regions as rows - are in our work-in-progress category and we hope to provide such functionality down the road. For the moment, the average values for each cluster (i.e. the values displayed as heatmaps in the cluster display panel), are output to a text file and you can use them for external plotting. Spark will have prompted you to specify an output directory. In that directory, in the "clusters" sub-directory, you will see a "clusters.values" file which contains these values.
Hope that helps.
Cydney
Thanks for your feedback and interest in Spark!
Both of your feature requests - (1) summary line plots for each cluster, and (2) sortable heatmaps displaying regions as rows - are in our work-in-progress category and we hope to provide such functionality down the road. For the moment, the average values for each cluster (i.e. the values displayed as heatmaps in the cluster display panel), are output to a text file and you can use them for external plotting. Spark will have prompted you to specify an output directory. In that directory, in the "clusters" sub-directory, you will see a "clusters.values" file which contains these values.
Hope that helps.
Cydney
Reply all
Reply to author
Forward
0 new messages