Skip to content
Valentin Gologuzov edited this page Jun 20, 2014 · 2 revisions

Actual task

31.1.2014

CBMS paper, miXGENE -- a web tool for integrative gene expression data analysis using prior knowledge

Valentin: 1. workflow automated execution implementation -- the last steps, 2. user's data import -- including phenotypes in external file and platform annotation, 3. SubAgg and SVDAgg implementation.

Matej: 1. 2 pictures that exemplify a particular workflow (2 different ways of CV meta-plugin), share with Valentin, 2. CBMS paper outline, collection of the earlier texts relevant to the topic, 3. a section on workflow management, 4. a section on GE classification with prior knowledge.

Next meeting: Fri 7th of February 10am

Previous tasks

6.12.2013

Valentin: 1. ongoing task: XGENE workflow -- implement and show a simple workflow that contains cross-validation, focus on workflow configuration tools (scopes, validity checks, block inputs and outputs, running workflow), 2. thesis draft: workflow management description + ongoing review of workflow management systems and miXGENE x XGENE

Matej: finish plugin overview page (provide links to Python/R libraries or actual implementations relevant to the individual plugins)

Next meeting: Tue 17th of December 1pm

15.11.2013

Valentin: 1. ongoing task: XGENE workflow management -- implementation of workflow system with functionality identical to the original XGENE (the first input block has already been done + infrastructure), 2. review: XGENE x miXGENE comparison (programming tools/platforms, functions) bioinformatics workflow management systems (overview, inspiration)

Matej: 1. specification of future diploma theses that develop current implmentation, 2. plugin overview page

Review draft: Fri 22th of November

Next meeting: Fri 29th of November 10am

20.9.2013

Globalni: migrovat gene set funkcni XGENE workflow s dvema platformami na miXGENE

Valentin: schopnost importovat GEO soft-format (zrejme knihovna BioPythonu), implementace workflow s pouzitim BroadInstitute definice gene sets, workflow: import dat, konverze identifikatoru, vypocet features odpovidajicich gene sets (prumerovani, PCA), analyza novych features pomoci t-testu (seradit features dle kvality), PCA (vizualizace vzorku v novem feature prostoru) a SVM (klasifikace)

Matej: doda BroadInstitute gene sets, nove plugins v R knihovne miXGENE

Do budoucna: import RNA-seq dat, nejlepe zahrnujici soucane mRNA a miRNA, vyuziti targets

Dalsi schuzka: patek 4.10. v 10h

11.9.2013

Matej: summarizes and delivers R packages that form a meaningful XGENE workflow Valentin: implements an import of mRNA GSE from NCBI, implements a miXGENE workflow using packages from Matej Jiri: checks whether we can import mRNA, miRNA and methylation NCBI data that match

checkpoint: Fri 20th 10am

Update 2013.10.25

Done:

  • We can merge msigdb gene sets with GSE datasets with PROBE ID as units
  • Splitting dataset into test and train sets
  • User is able to assign samples to user defined classes
  • Files fetched from NCBI are now cached
  • Update version of the R miXGENE package

Update 2013.10.11

Done:

  • Parsing SOFT format into probe expression table, probe <-> entrez_gene_id mapping.
  • Parsing and generation .gmt format, and basic support for gene sets.
  • PCA, SVM, Ttest and global test can be applied to arbitrary GSE (here is problem with sample classes, now we try to infer class from sample title as prefix ended by "_" or "," ). Small and good example GSE25724 used entrez_gene_id units.
  • Routine update in UI: now configuration parameters can be grouped.
  • [Bugfix:] Context update is now atomic action.

What's next: