Permuation-tail probability (PTP) test for an evolutionary signal
0
1
Entering edit mode
9.7 years ago

Recently I'd like to construct a supertree using protein sequences from different species. The pipeline has been given in http://mbe.oxfordjournals.org/content/23/1/74.full and http://www.biomedcentral.com/1471-2148/6/99 . In both papers, the authors used Permuation-tail probability (PTP) test to determine whether an evolutionary signal that was significantly better than random noise (P< 0.01). But there is little about how to perform PTP test. So could someone know how to perform PTP test?

PTP = the proportion of all data sets (actual plus randomized) producing a tree as short or shorter than the length for the actual tree.

Based on this paper, my understanding was: if I have a multiple alignment results of a group of proteins, 1) I first need to construct a tree which is the actual tree and calculate the tree length; 2) I need to define only conserved alignment blocks using Gblocks and do 99 times of permutation by reshuffling the conserved alignment blocks; 3) then I need to construct 99 phylogenetic trees based on the 99 datasets. 4) calculate PTP, we can say the phylogenetic signal is significant if PTP < 0.05.

The problem is 1) I don't know whether my understanding is right, 2) how to calculate a tree length given a tree file. Should I just add up the values in the tree? Like a tree file shown below:

#########################################

(
(
(
(
(
(
Chrysemys-picta-bellii_1_rep:0.13855,
Chrysemys-picta-bellii_8_rep:0.13614)
:0.30867,
Chrysemys-picta-bellii_9_rep:0.48048)
:0.00949,
(
Chrysemys-picta-bellii_3_rep:0.41386,
Chrysemys-picta-bellii_6_rep:0.41506)
:0.04593)
:0.00502,
Chrysemys-picta-bellii_4_rep:0.46094)
:0.00840,
Chrysemys-picta-bellii_2_rep:0.44416)
:0.00282,
Chrysemys-picta-bellii_5_rep:0.44868,
Chrysemys-picta-bellii_7_rep:0.44529);

#########################################

In the software called clann, it does include YAPTP (Yet another PTP) test. I think the differences is the null hypothyses which is phylogenetic signal in the actual tree was no better than random. So what's the equation to calculate YAPTP? I can program using R or Perl. So if you give me the equation, I use it to do calculation.

Thanks

alignment statistics phylogenetics • 2.5k views
ADD COMMENT
0
Entering edit mode

Can you be more specific? Both papers cite the orginal description of the specific test they use, and the MBE paper cites the software suite. Is there something specific you are having trouble with, given that information?

ADD REPLY
0
Entering edit mode

Hi David, thanks for your replying. I edited the original post. Is this clear? Best.

ADD REPLY

Login before adding your answer.

Traffic: 1724 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6