BIOSBRED@UMCVMB.MISSOURI.EDU ("FELIX BREDEN") (06/28/91)
- - The original note follows - - Somebody please tell me if this gets posted. I cut total genomic dna from guppies with HINF1 and ALU, and probed with the repeated sequences (GACA)4, one of the repeated sequences associated with sex chromosome of the heterogametic sex in higher vertebrates. Obtained RFLP patterns that are highly polymorphic, up to 30 bands for these populations, 16-20 per individual. I want to use this info for phylogenetic reconstruction. Using band sharing as a distance measure, build a phenogram, or maybe use presence of a band as a character and do a cladistic analysis. Aware of Lynch's 85 criticism of (1) nonlinearity of relatedness measures based on fingerprints and (2) high variance around estimate of relatedness. Regarding (2), given large number of bands, and ability to survey large number of offspring from lab matings within each population to eliminate comigrating bands, should be able to reduce variance. regarding (1), many phenotypes used to produce distance measures are probably not linear, so that shouldnt totally invalidate band-sharing data, and Lynch agrees that bandsharing should give rankorder relatedness, so why shouldnt it work in a cladistic analysis? Grant has to go in Monday, so any thoughts greatly appreciated. Felix Breden