search

galaxyproject / galaxy

Data intensive science for everyone.
https://galaxyproject.org
Other
1.41k stars 1.01k forks source link

2016 GCC Data Hackathon Hub #2520

Closed jennaj closed 8 years ago

jennaj commented 8 years ago

New Home: https://github.com/bxlab/galaxy_hackathon/issues/31

yvanlebras commented 8 years ago

Hi everyone,

A first rapid proposition list I wrote for myself and maybe can be of interest... or not ;) :

bgruening commented 8 years ago

@yvanlebras see this idea on the code side: https://github.com/bxlab/galaxy_hackathon/issues/4

yvanlebras commented 8 years ago

Thanks @bgruening ! Concerning interactive tour, I found the Bootstrap tool.. Thank you also for info!

griffinp commented 8 years ago

Philippa Griffin here... I would like to work on area 4) finding some good example datasets - e.g. for variant calling, and/or for RADseq (but happy to work on other topics) I will only be here part of Saturday as I'm teaching on Sunday.

nathanweeks commented 8 years ago

Ideas:

  1. Create a Galaxy tool for the Canu Assembler
  2. Pipelines in Galaxy (UI design and backend implementation w/ Unix pipes)
ghost commented 8 years ago

Ideas (Krzysztof Poterlowicz):

  1. Galaxy flavour for epigenomics analysis 2.Teaching materials for biomedical students (i.e epigenetics)
ssander5 commented 8 years ago

I am most interested in working on: 1) writing tutorials that go along with the Shared Data Libraries 2) helping set up and establish Shared Data Libraries 3) developing best practice workflows

Also, I don't know if this exists already (as I said, I am new to galaxy), but I think it would be useful to have some means of exporting workflows into command line syntax. My previous lab was half command line friendly and half not, both stubbornly sticking to what they knew. I don't think this is uncommon and it would be useful to have the ability to export a workflow used in galaxy to also be used on command line, with the same parameters, etc. This would also serve as an easy way to transition people from learning bioinformatics on galaxy to also gaining comfort/familiarity on command line. Just a thought ^_^.

MoHeydarian commented 8 years ago

Data visualization/plotting - the chart/visualization tools in Galaxy are limiting. It would be nice to be able to visualize/plot data with the same flexibility you have in R (with ggplot2, for example), but not have to move your data out of Galaxy.

devikaatgit commented 8 years ago

I have 3 points to make:

  1. I am interested in developing a stable pipeline/workflow for genomic as well as transcriptomic assembly and differential expression of prokaryotic microorganisms - for both paired end and single end reads... - easy to use for biologists like me
  2. Integration of bioconductor package workflows; I am specially interested in Ballgown - an R package for statistical analysis of assembled transcriptomes that I wish to use after Stringtie.
  3. Also i feel it would be good if the data visualization tools can be improved a little more, like plotting graphs and even networks (Genes as nodes/Interactions or co-expression data from RNA-seq etc as edges)... Even venn diagrams now take up only upto 3 sets... aspects like that can be improved...

My points come from the application point of view that i make use of in Galaxy. I am still not sure if these are the ones expected for the hackathon,

Nevertheless these are what I think of. Let us move forward.

Happy Hackathon everyone!!!

BoughAida commented 8 years ago

I am interested on:

  1. Creating Galaxy workflow for variant calling with GATK for prokaryotic genomes (existing already in command-line)
  2. Exporting a Galaxy workflow into command-lines

I would like to help on Finding some good example datasets - e.g. for variant calling (Yvan's idea)

firaan1 commented 8 years ago

Aarif here... I have a couple of issues and interests...

accessing galaxy from commandline api (sort of), handling dataset names in workflow and deleting intermediate datasets after workflow completion, rnaseq variant calling pipeline, ggploting tools Multiple history panel functionalities apply workflows directly on data from datalibraries ( not just histories) Openbis and galaxy (any hope??)

all other topics too

BoughAida commented 8 years ago

I am also interested on devikaatgit's points. We have the same research issues.

MoHeydarian commented 8 years ago

Agree with Aarif. Handling data set names in workflows needs improvement. Some tools cooperate with naming terminal data sets, but most do not.

jxtx commented 8 years ago

@MoHeydarian regarding dataset names, this is probably relevant: https://github.com/bxlab/galaxy_hackathon/issues/13

ghost commented 8 years ago

Hi Guys, I am interested in the couple of things you mentioned including:

MoHeydarian commented 8 years ago

We should ask the Coding hackers if anyone is interested in updating Trackster to output high resolution images from genomic regions (like a pdf). The UCSC genome browser does this, but this requires additionally formatting data (ex: BAM to bedgraph), renaming files, and moving it out of Galaxy (not ideal).

dannon commented 8 years ago

@MoHeydarian I know @msauria was interested in outputting nice trackster figures. See https://github.com/bxlab/galaxy_hackathon/issues/17

yvanlebras commented 8 years ago

Some interesting hackathon issues:

crutching commented 8 years ago

One thing we really need, and I think this need is universal, is a framework from which to validate genomics workflows quickly. As new tools are developed, it is difficult for clinical genomicists to actually adopt them as the barrier towards CLIA/CAP validation is significant. Additionally, finding the correct parameter sets for a given tool is also not trivial, and requires some sort of rapid parameter sweep and results evaluation.

kkamieniecka commented 8 years ago

Galaxy workflow for epigenomics analysis