In article <01HUHHX4FA2C00AWKF at vmsa.oac.uci.edu>, you say:
>>The best estimate of the intrinsic Ts/Tv rate is probably that
>determined by comparison of the most closely related sequences
>you have.
Is there any experimental support for this? References?
>The more distant the sequences, the more this rate
>is randomized.
In my dataset of molluscan 18S sequences, TS/TV ratios vary from 0.9 to 3.5.
The high values in the closely related species, but average of 1.4.
High TS/TV ratios add a lot of standard error/variation to branches, bootstraps
values correspondingly decrease.
So, is it really the best option to take the TS/TV ratio of the most closely
related species?
Gerhard Steiner
