|
The Backbone-Dependent Rotamer Library
PDB Chains Used
The 850 chains used in constructing the May 2002 version of the backbone-dependent
rotamer library are all from x-ray crystallographic
structures from the Protein Data Bank. The list was
obtained from our PISCES server. The maximum resolution of chains in the list is 1.7 angstroms.
Each chain in the list has less than 50% sequence identity with every other chain in the list.
The maximum R value for any structure in the list is 0.25.
The list of PDB chains is available:
bbdep02.May.cmpdlist.gz
Raw Data
Side-chains in the 850 PDB chains in the data set were further filtered using the criteria
of Lovell et al. (see References below)
First, hydrogen atoms were built with the
REDUCE program. REDUCE
adds hydrogen atoms and flips some asparagine, glutamine, and histidine side chains to satisfy
hydrogen bonding requirements. Side-chains of these types where the flip state was ambiguous
were removed from the data set used to build the library.
Second, side-chains with B-factors over 30.0 were excluded from the data set.
Third, the
PROBE program was used to identify side-chains with atom-atom clashes, including
hydrogen atoms. Side-chains with clashes were eliminated from the data set, according to
the criteria of Lovell et al.(see References)
The raw data used to build the libraries are available in the file
bbdep02.May.dihedrals.gz.
The format of the file is as follows:
Res iphipsr1r2r3r4 phi psi chi1 chi2 chi3 chi4 omega occ B-fac resno chain ss PDB
ARG 25 17 3 2 2 1 -110.58 166.39 289.39 185.52 175.07 119.23 179.52 1.000 23.86 9 A C 1DYSA
ARG 30 31 2 2 1 1 -58.59 -46.53 206.63 184.83 68.43 81.12 176.60 1.000 27.75 25 A H 1DYSA
ARG 24 1 3 2 3 3 -115.07 10.94 303.65 189.64 308.26 251.57 175.31 1.000 30.23 31 A C 1DYSA
ARG 30 33 1 2 1 2 -60.66 -28.48 71.00 195.69 54.87 212.64 176.96 1.000 24.06 60 A H 1DYSA
ARG 29 33 3 2 2 2 -67.84 -31.24 295.75 181.71 164.20 145.66 -179.98 1.000 22.53 70 A H 1DYSA
where
iph and ips are rounded off values of phi/10.0 and psi/10.0 respectively
with phi and psi expressed between 0 and 360 degrees
r1, r2, r3, and r4 are integers that describe the chi1, chi2, chi3, and chi4 rotamers according to Dunbrack and Cohen (see references)
with some modifications for Asn, Gln, and His
side-chain dihedral angles are given between 0 and 360
the highest B-factor for any atom in the side chain is given
the residue number (resno), chain, and PDB entry (with chain appended) are also given
ss is the secondary structure as given by the stride program of Frishman and Argos
Backbone-Independent Rotamer Library
The backbone-independent rotamer library refers to the individual populations
of rotamers, without reference to the local backbone conformation angles
(phi and psi).
The backbone-independent rotamer library is compiled using Bayesian
statistics (See paper by Dunbrack and Cohen in References).
This provides a more accurate estimate of the populations
of rare rotamers in an infinitely large dataset from the more limited dataset
of the PDB.
This version of the library differs also in that it contains conditional
probabilities of the chi2,chi3,chi4 rotamers based on the identity of the
chi1 rotamer. Since the local backbone conformation has such a strong effect
on the chi1 rotamer probabilities, we have derived the conditional backbone-independent
rotamer library.
As an example, a fragment of the backbone-independent rotamer library
for leucine is shown below. It contains the non-conditional rotamer library, p(r1,r2),
alongside the conditional rotamer library, p(r2|r1) where
"r1" refers to the chi1 rotamer and "r2" refers to the chi2 rotamer, and
sig is the sqrt(variance) of the previous column in each case:
Res r1r2r3r4 n(r1) n(r1,r2) p(r1,r2) sig p(r2|r1) sig chi1 sig chi2 sig
=================================================================================
LEU 1 1 0 0 148 97 0.73 0.06 65.54 3.18 58.7 14.5 80.7 17.6
LEU 1 2 0 0 148 48 0.36 0.04 32.43 3.13 71.8 13.9 164.6 14.8
LEU 1 3 0 0 148 3 0.02 0.01 2.03 0.94 58.2 13.7 -73.6 12.4
LEU 2 1 0 0 4222 3830 28.91 0.32 90.72 0.36 -179.2 9.9 63.0 8.1
LEU 2 2 0 0 4222 339 2.56 0.11 8.03 0.34 -164.9 17.4 170.7 27.0
LEU 2 3 0 0 4222 53 0.40 0.04 1.26 0.14 -169.8 14.2 -73.1 18.1
LEU 3 1 0 0 8877 483 3.65 0.13 5.44 0.20 -86.1 14.9 54.6 24.5
LEU 3 2 0 0 8877 8282 62.52 0.34 93.30 0.22 -64.7 9.1 174.5 8.6
LEU 3 3 0 0 8877 112 0.85 0.06 1.26 0.10 -81.5 12.8 -55.0 19.6
Download the file bbind02.May.lib.gz
To derive the backbone-independent rotamer library, we use the conditional
probabilities of each rotamer about each chi angle dependent on the previous
chi angle in the side chain. These data are available in the file bbind02.May.condlib.
As an example, we show the data for methionine:
Res ra= i rb= j n(i,j) p(j|i) chi1 sig chi2 sig
==============================================================
MET r1= 1 r2= 1 18 8.14 59.4 ( 15.4) 79.8 ( 10.1)
MET r1= 1 r2= 2 196 88.69 64.1 ( 8.8) -178.8 ( 11.4)
MET r1= 1 r2= 3 7 3.17 72.5 ( 13.4) -75.2 ( 15.8)
MET r1= 2 r2= 1 292 32.30 -175.3 ( 10.5) 66.2 ( 10.6)
MET r1= 2 r2= 2 550 60.84 -176.4 ( 11.9) 178.5 ( 11.7)
MET r1= 2 r2= 3 62 6.86 -178.7 ( 9.5) -82.4 ( 7.6)
MET r1= 3 r2= 1 25 1.29 -78.5 ( 14.7) 67.6 ( 15.9)
MET r1= 3 r2= 2 1146 59.22 -67.5 ( 9.2) -180.0 ( 11.5)
MET r1= 3 r2= 3 764 39.48 -65.3 ( 9.8) -62.5 ( 10.6)
MET r2= 1 r3= 1 240 71.64 64.9 ( 10.1) 72.6 ( 15.1)
MET r2= 1 r3= 2 64 19.10 70.7 ( 11.8) -170.7 ( 24.6)
MET r2= 1 r3= 3 31 9.25 75.8 ( 14.3) -97.4 ( 20.6)
MET r2= 2 r3= 1 792 41.86 178.7 ( 12.0) 71.4 ( 14.7)
MET r2= 2 r3= 2 443 23.41 178.5 ( 10.6) -178.5 ( 20.1)
MET r2= 2 r3= 3 657 34.73 -178.3 ( 11.5) -72.7 ( 14.9)
MET r2= 3 r3= 1 98 11.76 -64.9 ( 11.1) 99.6 ( 10.6)
MET r2= 3 r3= 2 112 13.45 -65.7 ( 13.3) 172.2 ( 24.0)
MET r2= 3 r3= 3 623 74.79 -63.7 ( 11.4) -69.9 ( 13.3)
Download the file bbind02.May.condlib.gz
Backbone-dependent Rotamer Library
The backbone-dependent rotamer library refers to the individual populations
of rotamers for small phi,psi bins. It is analogous to calculating
a Ponder & Richards type library for small regions (10 degrees x 10
degrees) of the Ramachandran map.
These two versions of the library give the expected frequencies for
all rotamers of all sidechains. For Arg and Lys, this means 81 rotamers
for each phi,psi bin. For Met, Glu, and Gln, this means 27 rotamers.
It should be noted that some of these rotamers are very rare and some have
never been observed. Use with caution. See the review by Dunbrack (2002).
This version of the library is very large. It now also contains the
standard deviations for each dihedral angle at each phi,psi.
There is a sorted version and unsorted version. Here are a few lines
of the unsorted version for Met:
Res phi psi N r1r2r3r4 p chi1 chi2 chi3 chi4 sig1 sig2 sig3 sig4
MET -80 -40 282 1 1 1 0 0.000051 62.7 80.1 72.5 0.0 12.6 8.6 13.3 0.0
MET -80 -40 282 1 1 2 0 0.000026 57.9 78.9 177.4 0.0 17.0 12.8 26.0 0.0
MET -80 -40 282 1 1 3 0 0.000002 59.4 77.8 -97.4 0.0 15.4 12.6 20.6 0.0
MET -80 -40 282 1 2 1 0 0.000341 66.8 -179.4 78.8 0.0 7.6 10.3 14.7 0.0
MET -80 -40 282 1 2 2 0 0.000194 64.9 -176.3 -175.1 0.0 6.8 7.2 18.2 0.0
MET -80 -40 282 1 2 3 0 0.000329 62.0 -178.4 -67.0 0.0 14.7 13.6 18.2 0.0
MET -80 -40 282 1 3 1 0 0.000001 72.5 -70.1 99.6 0.0 13.4 14.6 10.6 0.0
MET -80 -40 282 1 3 2 0 0.000004 72.2 -70.4 174.3 0.0 12.8 14.7 23.8 0.0
MET -80 -40 282 1 3 3 0 0.000025 72.8 -72.2 -71.5 0.0 13.2 15.4 11.5 0.0
MET -80 -40 282 2 1 1 0 0.111728 -176.2 61.1 72.4 0.0 7.1 6.6 13.3 0.0
MET -80 -40 282 2 1 2 0 0.019640 -174.4 67.4 -167.4 0.0 8.7 7.3 17.5 0.0
MET -80 -40 282 2 1 3 0 0.007668 -172.6 75.9 -103.2 0.0 9.4 10.0 16.7 0.0
MET -80 -40 282 2 2 1 0 0.041188 -177.7 -173.9 72.4 0.0 10.6 19.6 11.3 0.0
MET -80 -40 282 2 2 2 0 0.021308 -174.9 175.0 176.6 0.0 9.2 10.1 18.5 0.0
MET -80 -40 282 2 2 3 0 0.051293 -170.0 174.7 -70.9 0.0 16.2 9.5 17.8 0.0
MET -80 -40 282 2 3 1 0 0.000252 -178.7 -73.7 99.6 0.0 9.5 12.9 10.6 0.0
MET -80 -40 282 2 3 2 0 0.000980 -180.0 -79.5 173.2 0.0 13.4 11.7 20.4 0.0
MET -80 -40 282 2 3 3 0 0.009927 -179.7 -83.3 -74.8 0.0 6.2 5.7 9.3 0.0
MET -80 -40 282 3 1 1 0 0.002134 -78.4 67.5 74.6 0.0 13.9 15.5 18.5 0.0
MET -80 -40 282 3 1 2 0 0.000690 -79.2 69.2 177.4 0.0 12.3 12.3 25.3 0.0
MET -80 -40 282 3 1 3 0 0.004988 -76.3 68.5 -101.9 0.0 10.9 11.9 15.7 0.0
MET -80 -40 282 3 2 1 0 0.170387 -66.8 -179.8 72.0 0.0 8.5 14.7 9.2 0.0
MET -80 -40 282 3 2 2 0 0.124948 -69.3 -179.8 176.9 0.0 7.4 10.3 24.1 0.0
MET -80 -40 282 3 2 3 0 0.124929 -68.9 -172.4 -73.8 0.0 6.5 6.0 10.2 0.0
MET -80 -40 282 3 3 1 0 0.087400 -66.3 -62.9 96.1 0.0 5.4 6.6 11.5 0.0
MET -80 -40 282 3 3 2 0 0.026709 -64.5 -64.2 178.4 0.0 6.1 10.6 26.0 0.0
MET -80 -40 282 3 3 3 0 0.192860 -67.8 -60.2 -68.7 0.0 8.2 11.8 11.6 0.0
where phi and psi are the centers of 10x10 degree bins, N is the number of Met side chains
in this bin in the data set, p is the probability for the rotamer conformation given by
r1,r2,r3,r4, with dihedral angles chi1,etc. with standard deviations sig1, etc.
Download bbdep02.May.lib.gz - Backbone-dependent rotamer library
Downlaod bbdep02.May.sortlib.gz - Backbone-dependent rotamer library sorted by probability
References
R. L. Dunbrack, Jr. Rotamer libraries in the 21st century. Curr. Opin. Struct. Biol.
12, 431-440 (2002).
R. L. Dunbrack, Jr. and F. E. Cohen. "Bayesian statistical analysis of
protein sidechain rotamer preferences ." Protein Science 6,
1661-1681 (1997).
R. L. Dunbrack, Jr. and M. Karplus. Backbone-dependent Rotamer Library
for Proteins: Application to Side-chain prediction. J. Mol. Biol.
230, 543-574 (1993).
R. L. Dunbrack, Jr. Conformational analysis of protein sidechains: Empirical
energy parameters for proline and development of a backbone-dependent rotamer
library. Ph. D. dissertation, Harvard University (1993).
R. L. Dunbrack, Jr. and M. Karplus. "Conformational analysis of the backbone-dependent
rotamer preferences of protein sidechains." Nature Struct. Biol.
1, 334-340 (1994).
|
