Dear morphometricians, I am wondering if you could give me some suggestions on NTSYS output results. I have applied PGLS (Phylogenetic Generalized Least Square) in order to test for correlation between Centroid Size and shape variables (partial warps and uniform components) under phylogenetic control. The tree used (a nexus file created without a character matrix but just from published trees) includes extant and extinct species and as branch lenghts (BL) represent time of divergences not all species are at the same tip level. Without Phylogenetic control a multiple regression is significant between Log CS and shape data (Wilk's lambda equal to 0.38, Probability = 6.97*10^-7) but when I include the OTU_covariance matrix (from the tree with BL in Ma) Wilk's lambda has an high value (67.77) and Probability is equal to 99.000 (I suppose it's a message for non-significant). When the OTU covariance matrix is scaled to unity (in my case 58 Ma = 1) Wilk's lambda is highly significant (0.000136 with Probability equal to 7.18*10^-143!). What about the correct interpretation? Are there any assumptions concerning the branch lenghts in PGLS? I have also randomized the Log Centroid Size data and the results still to be higly significant when I include a OTU covariance matrix scaled to unity. Non-transformed Centroid Size didn't change results. Thank you in advance for your suggestions Carlo Meloro
Carlo Meloro Phd student in Earth Science Dipartimento di Scienze della Terra Università degli studi di Napoli "Federico II" L.go San Marcellino 10 80138, Napoli, Italy phone: ++39 0812538331 e-mail: [EMAIL PROTECTED] ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program. -- Replies will be sent to the list. For more information visit http://www.morphometrics.org
