Skip to content
Records of my visit to the Dunn Lab at Brown University Winter 2014.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
acoel_phylogenomics
priapulid_profiling
README.md
calendar.md

README.md

bcv-at-brown

Records of my visit to the Dunn Lab at Brown University Winter 2014.

Gene expression profiling of priapulid development

Objective: characterize differentially expressed genes during the development of a priapulid worm using RNAseq data. Read more.

Progress:

  • Fixed Fritzen's installation quirks for Agalma.
  • Assembled new reference with Agalma 0.3.5.
  • Prepared data, calculated read counts, replicate plots.
  • Generated STEM profiles.
  • Ran differential expression analysis and got significant change in expression for many genes that correlate with developmental events.
  • Started playing with visualizing the data.

Plan:

  • Consolidate data analyses and differentially expressed genes.
    • Review RNAseq literature and get to know all the caveats of RNAseq and differential expression analyses.
    • Thoroughly understand the statistical strength and weaknesses of the data, including all implicit assumptions.
    • Delineate the assumptions of the data and how the differential expression test is being executed.
    • Try different DE methods.
  • Annotate genes with Gene Ontology and merge with results.
  • Select relevant developmental pathways and describe gene activity.
  • Explore the possibility of comparing with data sets of other species.
  • Think about data visualization.

Phylogenomics of Acoelomorpha

Objective: solve the phylogenetic position of acoels. Read more.

Progress:

  • Developed a script/module to search and filter SRA records.
  • Searched and got a list of SRA packages.
  • Tested workflow with a few organisms.
  • Defined curation criteria: if unique, yes; if multiple, pick mixed tissues; if tissue specific, merge multiple until 40M reads.
  • Running assemblies.

Plan:

  • Continue building assemblies.
  • Solve issues like quality offset, no rRNA, too large samples, multiple samples.
  • Define list of additional RNAseq assemblies.
  • Define list of additional genomic assemblies.
  • Study phylogenetics.
You can’t perform that action at this time.