User:R. Eric Collins/MBL/Edwards

• most common gene tree can not match the species tree (laura k mentioned this 'anomaly zone')
• likelihood methods based on coalescence can correct for this given many loci
• in this case concatenation guarantees wrong tree given lots of data

using full set of good trees rather than just best

• treeviz: tree set visualization projecting tree set into MDS space
• concaterpillar: measure of conflict in dataset

• incomplete lineage sorting methods
  • BEST, STAR, BUCKy, concatenation

• concatenation gives higher 'confidence' but how confident are you in the confidence?

• eigensoft: hetero/homozygous matrix of alleles

• royal ontario museum
  • museums are natural places for genome sequencing

• bayestraits: analyse morphological characteristics with sequence data
• small genomes may have first arisen in early bird ancestors, much before origin of flight

• large populations --> small genomes
• allow natural selection to work unimpeded by genetic drift
  • small deletions generally kept whereas in small populations they could get fixed easier
  • but what about protists with giant genomes? e.g. diatoms

• base frequency patterns
  • e.g. 5-letter words ending in CG depauperate since CG easily mutates to something else
  • even after correcting for GC content
  • word frequency --> "Bayesian Surprise" (Yogesh)

• Mycoplasma gallisepticum 1Mb
  • 2% of genetic diversity from M. gal. population made it into House Finch strain (using MIGRATE)

• expanding populations can have long tips and short internal branches during expansion
  • 17 sequenced strains of Mgal showed variations in population size and evolutionary rate

• CRISPRs as Lamarckian evolution...
  • time-sampled strains show rapid evolution of CRISPR loci

Questions

• can you use microbial flora/parasites/pathogens to delineate species?
  • baross did this with bacteria/phage