Moore Notes 7 28 10

Group Call

• Kelly visiting UCD today

• Steve's PLoS Bio paper
  • JG made major edits
    • streamlining to remove repeated ideas
  • SK working on it this afternoon, close to submitting
  • Cover letter: say that they talked about in the letter, spell out responses without repeating

• Tom's NSF fellowship issues

• JG, AD, TS, and JL will be at ISME

• SK and JL are going to ESA

• Josh - Saliva data project slides
  • Nasidze et al (2009) Genome Research saliva microbiome data
  • TS ran ESPRIT to get OTUs
  • JL's results about relationship between distance decay and ranges (area, perimeter, angularity)
  • Applied model to saliva data
  • Ranges can be in geographic or niche space
  • Niche data from WHO database: e.g., drinking water, dentistry, tobacco
  • Issue: niche data is national (or city), OTUs are on individuals
    • Lumping all OTUs for a country - 12 data points
    • SK: try contacting Nasidze et al to see if other data was collected
  • In niche space, have many dimensions.
    • How many degrees of freedom (unique directions) - need math for >2 dimensions
    • Could do PCA to see directions of greatest variability
    • Which pairs of niche data produce random sampling (not too correlated with each other)?
  • Some variables may be irrelevant or have too little data
    • adolescents vs. adults (check with Nasidze)
    • rural vs. urban (all urban saliva)
  • May be hard to make actual range maps
    • other good methods already exist, e.g. variations on convex hulls
    • only have 12 data points
  • Isoperimetric quotient (roundness of range) is measured well
    • important property of range shape
    • could be used to say something about dispersal properties
    • JG: but is not intuitive - map better
    • possibly constrain range methods using estimated area, perimeter etc (mean, sd of these)
  • Empirical distribution function of isoperimetric quotient
    • summary of histogram across all OTUs
    • drinking water x population: most OTUs have elongated ranges
    • most EDFs jump up around 0.2, but some are flatter
    • differences between EDFs probably more meaningful that absolute values
    • look at rank-abundance curves of OTUs: is it being driven by lots of rare OTUs?
    • try varying cutoff for OTUs in ESPRIT
  • Would be nice to include dynamical, dispersal model
  • Visually, how do ecologists display niche data typically?
    • heat maps on geographic maps
    • Michael Gasner (SFI) has method that stretches maps based on meta-data
    • Mark Newman - Atlas of the Real World
  • Other data sets? What in paper?
    • Saliva: this and other paper (PCR-based)
    • Global rRNA: status?
    • HMP mouth data