Does resampling change tree topology?

[Jakob Jilg]

Hi all,

I try tro write TNT scripts. What I basically do is tree search -> strict consensus -> resampling with bootstrap.

I think my script is saving one tree after strict consensus, and one after resampling. I'd expect them to have the same topology, with the consensus tree without support values for nodes and the tree saved after resampling with support values.

However: The trees have different topology. The consensus tree has many polytomies (as expected), while the bootstrap tree after resampling is much better resolved. 

xwipe;
macro =;                                                                   /* Activate Macro language to allow comments.         */
log results/log.txt;                                                       /* keep log of analysis */
mxram 1000;                                                                /* set ram to 1gb */

taxname = ;                                                                /* use taxon names for terminal nodes */
nstates NOGAPS;                                                            /* treat gaps as missing data as opposed to an additional state */
proc data.tnt;                                /* read matrix */
outgroup Cassida_nebulosa;                                                 /* define outgroup taxon */
hold 100000 ;
log results/tnt_run_log ;
taxname = ;
hold 10000; xmult=ratchet 1000 drift 100 fuse 100 hold 10000; bbreak=tbr;
nelsen * ;
tchoose { strict } ;
ttags = ;
blength *;
ttags );
export - results/consensus_tree.tre ;
ttags -;
ttags = ;
collapse tbr ;
resample replications 1000 ;
keep 1 ;
ttags );
export - results/boots.tre

ttags &nelsen_bootstrap.svg;

proc/;

This script is modified from https://isu-molphyl.github.io/EEOB563/computer_labs/lab2/TNT.html

All the best,

Jakob