Names of species/branches: species A, B, C, and D
can be replaced by authentic names.
Note: names should not contain spaces or commas inside (e.g., “gray
wolf” must be written as “gray_wolf”).
Settings:
Algorithms: Reverse is the preferred
algorithm.
In the case of rapidly radiating species, the Stepwise
algorithm might provide a better resolution.
Significance level: Default level “0.05” (95%
probability) gives a balanced resolution for most cases, and level
“0.01” (99% probability) rejects complex trees in favor of more simple
trees in complex cases.
It could be used in case of contradiction between chi-square results and
Empirical distribution or in case of not stable results of Empirical
distribution.
Criterion (Selection of approximations): The
Chi-square criterion allows getting estimations in a few
seconds.
However, Chi-square could lead to an incorrectly preferred
polytomy at the resolution boundary.
The Empirical distribution processes via a bootstrapping-like
procedure to calculate critical values for each comparison and give a
better resolution.
With an Empirical sample, size scale users could
stabilize unstable results of Empirical distribution runs by
more exact calculations of critical values.
However, the sample size is directly proportional to the running time of
the Empirical distribution criterion (e.g., with a sample size equal to
500, a run takes tens of minutes, and with a sample size of 2000, it
takes hours).
Marker Input: For the four lineages A,B,C,D, ten
different patterns of relationships (y11-y44) can be derived.
The supportive number of verified diagnostic TEs of each pattern has to
be inserted in the Marker Input field.
For calculation, the minimum input is one marker.
File Upload: Alternatively, all data can be uploaded from a text or excel file corresponding to the following example
With Start hammlet calculations! the process is
commenced.
With Reset the user can return all values and settings to
initial/default values.
Example:
Names of species/branches: human, chimpanzee, gorilla and orangutan.
Algorithm: Reverse (stringent)
Significance level: 0.05 (default)
Criterion: Empirical distribution (exact and slow)
Empirical sample size: 1000
Marker Input: 0 0 0 0 0 0 1 0 37 48
The resulting Most Probable Tree Topology
is indicated as T2 for the exemplified data, and the tree’s significance
is placed as text below the tree topology.
KKSC values for individual tree splits are given in the grey part.
The log-likelihood table (LL-table) can be downloaded and provides
optimal parameters for all five models.
With Download LL table (tsv format), one can retrieve the
result table of the best trees from all five groups for further
investigations.
Model indicating the tested model.
Alias describes models in “TTgg”
annotation;
Order provides information about the
permutation (where 1234 is equivalent to A,B,C,D);
LL presents the log-likelihood value for the
resulting trees;
Ratio presents the value of the double
log-likelihood ratio between resulting tree and polytomy (reflecting χ^2
value);
T(1) and T(3)
provides the relative length of tree branches (T(1) and T(3)) that is
proportional to the number of generations on corresponding
branches);
γ1 and γ3 provide
values of the hybridization coefficients.