Ren,
have a look at
http://www.embl-heidelberg.de/obstruct/obstruct_info.html.
This has precalculated datasets at various levels of sequence id
(or you can submit a personalised query).
Mike
---------------------------------------
ctzhang at tju.edu.cn wrote:
>> Dear netters:
>> I'd like to know a protein database, satisfying the following conditions:
>> 1) The pairwise sequence identity less than 30%
> (i.e., non-homologous proteins);
> 2) Reflecting the current state of the Brookhaven Protein Data Bank.
>> Thank you in advance for sending me the relevant information. The more
> detailed, the better.
>> My e-mail address: ctzhang at tju.edu.cn>> Ren Zhang
--