I would like to use a specific library for design using new distributions of P (Res | phi, psi). I found 7 files in the rosetta_database directory that seem to be exactly what I'm looking for: Paa, P_AA, Paa_n, P_AA_n, Paa_pp, P_AA_pp, P_AA_pp_08.2009
Our lab is in the process of becoming a Rosetta developer officially once again, but that will take some time.
First, can anyone point me in the right direction as to the format of each of these files/how each of them are used?. Some, like P_AA seem fairly obvious, but some of the others I have no idea.
Second, are these all the files I would have to edit, or are there more?
Any help would be appreciated. Thanks!
Which lab are you? If you've got access, it's much more productive to ask on the developers' mailing list (where at least you get more people reading it than me and a few others).
I've forwarded the actual question to someone who knows more than me.
In standard rosetta runs, only the P_AA_pp file is read as input. If you include the "-correct" flag, then instead the P_AA_pp_08.2009 file is read. You can specify an alternate file to read with the "-corrections::score::p_aa_pp " flag.
The format for this file is non-obvious and most readily understood from the code that reads the file. Look in src/core/scoring/P_AA.cc at the function beginning on line 157. In particular, lines 188/189 actually read from the input stream. The format is something like:
PHI PSI AA3LETTERCODE COUNTS COUNTSTOTAL PROBABILITY
where "COUNTS" and "COUNTSTOTAL" both end up getting ignored by the code. The column widths are fixed.
4 characters for phi,
4 characters for psi
3 characters for the aa 3letter code
7 characters for the probability
Hope this helps.
That will help immensely! Thanks!
I'm in Roland Dunbrack's. We used to have access, but for some reason the username/pass doesn't work anymore.
I've sent an email to the email account you registered this board account to describing how to get SVN access.