BetaSearch is a fast method for indexing and querying beta-sheets for substructures composed of peptide- or hydrogen-bonded residues. Beta-sheets are indexed as beta-matrices, which are projections of beta-sheets onto a matrix of amino acid characters.
For example, the beta-sheet of ubiquitin:
will be projected to the beta-matrix (see "Generate beta-matrices from PDB files"):
...TITLE
..TKVFIQ
LVLHLT..
QRLIF...
...QK...
which can be queried for the following substructure (see "Query an index for beta-residue motifs"):
I..
VFI
.T.
which is found in the top three rows of the beta-matrix.
You can try BetaSearch for yourself via the web-interface at
First, clone this repository:
git clone git@github.com:kianho/betasearch.git
then install the dependencies below.
BetaSearch was developed for Python 2.7 using the following libraries:
- NumPy
- BioPython
- Whoosh
- NetworkX
- docopt
- ptgraph2 (included in this repository)
which can be installed in Ubuntu using the following commands:
sudo apt-get install python-numpy python-biopython
sudo apt-get install python-whoosh python-networkx python-docopt
or via the Anaconda Python Distribution:
conda install numpy biopython matplotlib whoosh pip
pip install docopt
BetaSearch uses dssp to obtain the secondary structure assignments of protein
structures. The DSSP executable needs to be installed into your system path
and renamed to dssp, it can be downloaded from:
- ftp://ftp.cmbi.ru.nl/pub/software/dssp/
The bin/make_bmats.py script is used to generate
beta-matrices from PDB files using the following single-line representation,
with fields separated by the ^ character:
ls -1 /path/to/pdb/*.{pdb,ent,pdb.gz,ent.gz} | \
python ./bin/make_bmats.py -o ./bmats.txt
output (bmats.txt):
...
sheet-3cd4-A-003^4^5^4^f^f^15^t cell surface glycoprotein cd4^human^homo sapiens^KKVEF,QLVTC,.SVQ.,..GQ.
sheet-3cd4-A-002^2^2^2^f^f^4^t cell surface glycoprotein cd4^human^homo sapiens^LL,VL
sheet-3cd4-A-001^3^3^3^f^f^8^t cell surface glycoprotein cd4^human^homo sapiens^VEL,IIL,AD.
sheet-3cd4-A-000^6^18^8^t^f^51^t cell surface glycoprotein cd4^human^homo sapiens^......KVVLGK......,.QKE-EVQLLVFGLTA..,.VEC-IYTD...ELTLTL,.FHW-KN.......TLSV,QNGLIK............,FLT...............
sheet-3cd4-A-004^2^2^2^f^f^4^t cell surface glycoprotein cd4^human^homo sapiens^GT,ID
sheet-1ubq-A-000^5^8^5^f^f^24^ubiquitin^human^homo sapiens^...TITLE,..TKVFIQ,LVLHLT..,QRLIF...,...QK...
sheet-1fyt-A-002^4^7^4^f^f^22^hla class ii histocompatibility antigen, dr alpha chain^human^homo sapiens^.LLKHWE,EVRCDYV,NVTWLR.,...VPK.
sheet-1fyt-A-001^4^10^5^t^f^29^hla class ii histocompatibility antigen, dr alpha chain^human^homo sapiens^..EVTVLT..,FKDIFCILVN,F-RKFHYLPF,P-L..ES...
sheet-1brp-A-000^16^33^9^t^t^152^retinol binding protein^human^homo sapiens^.....................CAD-SYSF-VFS,.....................L-LRCSYQ-VAY,......................GNDDHWIVDT.,....................GWYKMKFK.....,...............DVCADMVGTFT.......,...............RVRGKATASM........,..............R-L..VAEFSV........,.............KK-AMAYWTG..........,.......CAD-SYSF-V-FS.............,.......L-LRCSYQ-V-AY.............,........GNDDHWIVD-T..............,......GWYKMKFK...................,.DVCADMVGTFT.....................,.RVRGKATASM......................,.RL..VAEFSV......................,KKAMAYWTG........................
sheet-1mtp-A-001^6^15^6^f^f^62^serine proteinase inhibitor (serpin), chain a^N/A^thermobifida fusca^.FELTQPHQ......,DVRLRAQHIVTDV-Y,GAEGAAATAAM-M-L,RAKAWLANTLI-ARL,...LASRTVLW-V..,........DVR-T..
...
where the fields are (in left-to-right order):
- sheet identifier in the format:
sheet-<pdb id>-<chain id>-<ordinal sheet number> - number of rows in the beta-matrix
- number of columns in the beta-matrix
- number of beta-strands in the beta-sheet
followed by meta-data obtained from PDB header contents:
- beta-sheet bifurcation:
tif the beta-sheet is bifurcated,fotherwise - number of residues in the beta-sheet
- molecule name
- organism name (common)
- organism name (scientific)
finally, comma-separated strings denoting the rows of the beta-matrix. For example, the beta-matrix representation of the beta-matrix in ubiquitin (PDB ID: 1ubq) is:
...TITLE
..TKVFIQ
LVLHLT..
QRLIF...
...QK...
which has a corresponding single-line representation:
...TITLE,..TKVFIQ,LVLHLT..,QRLIF...,...QK...
The bin/make_index.py script is used to create a BetaSearch
index inside as a single directory from a file containing beta-matrices (see
bmats.txt output above):
python bin/make_index.py -d ./betasearch-index < ./bmats.txt
The bin/run_queries.py script is used to run queries on a BetaSearch index:
# run a single query from stdin
echo "query-XXX V.,LR" | python bin/run_queries.py -d ./betasearch-index
# run multiple queries from a file
python bin/run_queries.py -d ./betasearch-index -q ./queries.txt
where each query is formatted as a single line:
<query id> <one-line beta-matrix representation>
for example:
query-003 I..,VFI,.T.
queries for the motif:
I..
VFI
.T.
Multiple queries can be performed by appending more single-line queries.
Queries must:
- be composed of three or more connected residues,
- contain only residues from the 20 standard amino acids,
- form a rectangular matrix, padded using the "
." character.
For example:
IVK IV I MI. .K..
.K V .KE MIKE
K .A..
.N..
Examples of invalid queries:
IV IV IVL.
. ....
K KIAN
(< 3 residues) (disconnected residues)
..R.
KIAN
..O.
("R" is not one of the 20 standard amino acids)
KI
.AN.
..AIK
(non-rectangular, missing "." padding in the
first and second rows)
Please consider citing our paper if you find BetaSearch useful in your research:
- H. K. Ho, G. Gange, M. J. Kuiper, and K. Ramamohanarao, “BetaSearch: a new method for querying beta-residue motifs,” BMC Research Notes, vol. 5, 2012. [article]
see CONTRIBUTORS.md