Dear List, me and my collegues are trying to sort out the following problem: we have some trout samples from an italian lake where two different taxa are known to coexist. The results of a genetic model based clustering (STRUCTURE 2.1) suggest the exsistence of two well defined groups (the ancestral pool?) and of some intermediate (hybrid?) individuals. The algorithm actually provides an estimate of the ancestry of each individual from two differentiated gene pools. The two coexisting taxa are usually easily distinguishable by external human observation, though many uncertain individuals are found.
We would like to use the shape analysis tools (geometric morphometric or elliptic fourier analysis) in order to investigate and eventually describe the shape differences among the two taxa and the hybrid individuals.... but we would like to carry out a "blind" analysis, with no prior grouping. Substantially we don't want to make any assumptions about the belonging of individuals to any of the taxa. In this way, we hope to use shape analysis as a pattern recognition technique, comparing the pattern obtained with those from genetics. The question is: could some morphological groups be detected ? Can we arrange this? How? Our aim is to indipendently (genetic versus shape analysis) investigate the pattern generated from the two taxa and their hybrid. We would appreciate suggestions or "methodological corrections" from list.... ******************************* Tommaso Russo Ph.D. Experimental Ecology and Aquaculture Laboratory Biology Department University of Rome "Tor Vergata" Via della Ricerca Scientifica 00133 Rome (Italy) Tel +390672595972 Fax +390672595965 Mobile +393290734943 ******************************* -- Replies will be sent to the list. For more information visit http://www.morphometrics.org
