KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#developer-tools #typescript #dsh #dsh-plugin #ai-agent #codex #ai-agents #deepseek-harness #claude-code #cli #react #mcp

mmCIF2BioLiP

★ 34 Open GitHub ↗

BioLiP2 database curation and web interface

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# BioLiP: structure database for biologically relevant ligand-protein interactions #
This is the source code for curating a local copy of the BioLiP database from mmCIF files of the PDB database and displaying the database content in web browser.

## Installation for database update ##
```bash
cd script/
make
```
The compilation requires GCC with C++ 11 support.
The binary executable of [cd-hit](https://github.com/weizhongli/cdhit), [CSSR](https://github.com/pylelab/CSSR), and 
[US-align](https://github.com/pylelab/USalign) are available at ``script/`` for 64 bit Linux. 
When using other operating system, they can be recompiled:
```bash
# install script/cd-hit
cd cdhit/
make install
cd ..

# install script/CSSR
cd CSSR/
make install
cd ..

# install script/USalign script/pdb2xyz script/pdb2fasta script/xyz_sfetch
cd USalign
make install
cd ..
```
Additionally, we make the following binary executables available under ``script`` that should be reinstalled if not using 64 bit Linux:
* ``makeblastdb``, ``blastn`` and ``blastp`` from the [NCBI BLAST+](https://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/LATEST/)
* ``foldseek`` from [foldseek](https://mmseqs.com/foldseek/foldseek-linux-sse41.tar.gz)

## Installation for web browsing ##

1. This repository contains a copy of [JSmol](https://wiki.jmol.org/index.php/JSmol) under the ``jsmol`` folder. This is not needed for database update. It is solely for visualization through the web browser.

2. Optionally, to show directed acylic graph for Gene Ontology terms on the website, the ``dot`` program from [Graphviz](https://graphviz.org/) can be used. If your system does not have Graphviz, you can install it to ``graphviz/bin/dot`` from source code:
```bash
wget https://gitlab.com/api/v4/projects/4207231/packages/generic/graphviz-releases/6.0.2/graphviz-6.0.2.tar.gz
tar -xvf graphviz-6.0.2.tar.gz
mkdir ../graphviz
./configure --prefix=`readlink -e ../graphviz`
make
make install
rm graphviz-6.0.2.tar.gz
```

3. Follow ``output/readme.sh`` to set up for web browsing.

## Database update ##

The first time to set up the database takes at least two days. Subsequent weekly updates only take a couple hours. The database would requires ~100GB of disk space, including ~60GB to store original mmCIF files downloaded from PDB (``pdb/data/structures/divided/mmCIF/``; see Step 11 to reduce this requirement), ~20GB to store intermediate PDB files (``interim/``), and another ~20GB to store final files (``weekly/``, ``data/`` and ``download/``).

### Step 1: Download PDB entries ###
Run the following shell script the first time the database is created.
```bash
script/rsyncPDB.sh
```
This script downloads the initial set of mmCIF files for full asymetric unit to ``pdb/data/structures/divided/mmCIF``. It will also download the resolution of pdb to ``pdb/derived_data/index/resolu.idx``.

Alternatively, run the following script to download only the missing mmcif files and resolu.idx.
```bash
script/download_pdb.pl
```
Both ``script/rsyncPDB.sh`` and ``script/download_pdb.pl`` have the same purpose of downloading mmcif files and resolu.idx. For the first time the database is setup, ``script/rsyncPDB.sh`` must be used instead of ``script/download_pdb.pl`` because the later is much slower when a large number of files need be downloaded. For subsequent weekly update, either one of the script can be used. If you do not want to clean up the pdb/ folder, the first script ``script/rsyncPDB.sh`` is preferred. If you do want to clean up the pdb/ folder after every update, which can save ~60GB of disk space, you must use the second script ``script/download_pdb.pl`` to avoid re-downloading the whole pdb database every week. Hard disk space permitted, it is recommended to use ``script/rsyncPDB.sh``, which offers faster download speed and less file corruption.

### Step 2: Download SIFTS function annotation ###
This step can be performed in parallel to step 1.
```bash
./script/download_sifts.pl
```
This script downloads taxonomy, uniprot accession, EC, GO and pubmed ID to ``sifts/flatfiles/tsv`` and extract the data to ``data/*.tsv.gz``.
This script downloads gene ontology to ``obo/go/go-basic.obo``.
This script downloads swissprot to ``uniprot/current_release/knowledgebase/complete/uniprot_sprot.fasta.gz``.

Optionally, run the following script to download csa. The manually curated dataset of csa is updated infrequently. Therefore, it is not necessary to run it every week.
```bash
./script/download_csa.pl
```
This script downloads catalytic site to ``m-csa/api/residues.json`` and extract the summary to ``data/csa.tsv.gz``.

### Step 3: Download binding affinity ###
This step must be performed after step 2 (regardless of step 1) because it requires the pdb to uniprot mapping file generated in the previous step.
```bash
./script/download_bind.pl
```
This script downloads BindingDB to ``bind/BindingDB*`` and extract the summary to ``data/BindingDB.tsv.gz``.
This script may have the warning: "gzip: bind/BindingDB_All_tsv.zip: invalid compressed data--length error", which is caused by decompressing zip with zcat. This warning can be ignored.

Optionally, run the following script to download MOAD. Since MOAD only updates every few years, it is not necessary to run it weekly.
```bash
./script/download_moad.pl
```
This script downloads MOAD to ``bind/every.csv`` and extract the summary to ``data/moad.tsv``.

Again optionally, follow instructions at [bind/README.md](bind/README.md) for how to semi-manually create PDBbind-CN summary at ``data/PDBbind.tsv``. This only need to be peformed annually.

### Step 4: download auxilary data ###
This step can be performed independent of step 1 to 8, but must be done before step 9.
```bash
./script/download_ligand.pl
./script/curate_smiles.pl
```
This script downloads CCD ligand to ``pdb/data/monomers/components.cif.gz`` and extract their summary to ``data/ligand.tsv.gz``.
This script downloads enzyme to ``enzyme/enzyme.dat`` and extract their summary to ``data/enzyme.tsv.gz``.

### Step 5: convert mmCIF to PDB format ###
This step must be performed after step 1 (regardless of step 2 to 4).
```bash
./script/curate_pdb.pl
```
This script converts mmCIF files from ``pdb/data/structures/divided/mmCIF`` to ``interim/*/*.tar.gz`` and ``interim/*/*.txt``

### Step 6: download pubmed abstract ###
This step must be performed after step 2 and 5 (regardless of step 3 and 4).
```bash
./script/download_pubmed.pl
```
This script checks ``interim/*/*.txt`` for artifact ligand and download pubmed abstract to ``pubmed/*.txt``

### Step 7: remove artifact and unbound ligand ###
This step must be performed after step 6.
```bash
./script/curate_ligand.pl
```
This script check potential artifact ligand listed by ``interim/*/*.txt`` against pubmed abstract at ``pubmed/*.txt``.
It repackages receptor and biologically relevent ligand pdb files into ``interim/*/*.tar.bz2``. 
Binding sites are written to ``interim/*/*.bsr``.

### Step 8: prepare redundant dataset ###
This step must be performed after step 7.
```bash
./script/make_weekly.pl
```
This script reads ``interim/*/*.tar.bz2`` and ``interim/*/*.bsr`` and writes ``weekly/receptor_*.tar.bz2``, ``weekly/receptor1_*.tar.bz2``, ``weekly/ligand_*.tar.bz2`` and ``weekly/BioLiP_*.bsr.gz``.
FASTA sequence of protein, peptide, rna, dna are saved to ``data/*.fasta.gz``

### Step 9: prepare nonredundant dataset ###
This step must be performed after step 3, 4 and 8.
```bash
./script/make_nr.pl
```
This script converts ``weekly/BioLiP_*.bsr.gz`` to ``weekly/BioLiP_*.txt`` and ``weekly/BioLiP_*_nr.txt``.
It also creates ``weekly/receptor_*_nr.tar.bz2``, ``weekly/receptor1_*_nr.tar.bz2`` and ``weekly/ligand_*_nr.tar.bz2`` from intermediate files of the previous step.

### Step 10: assign RNA secondary structure ###
This step must be performed after step 8
```bash
./script/make_rSS.pl
```
This script assigns RNA secondary structure for ``weekly/ligand_*.tar.bz2`` and store it at ``data/rna_ss.txt.gz``.

### Step 11: curate EC annotation ###
This step must be performed after step 8.
```bash
./script/make_EC.pl
```
This script reads EC from ``data/chain2ec.tsv.gz`` and extract the summary to ``data/ec_all.tsv.gz``. It also creates ``weekly/Enzyme_*.tar.bz2``.

### Step 12: curate GO annotation ###
This step must be performed after step 11.
```bash
./script/curate_GO.pl
```
This script reads GO from ``obo/go/go-basic.obo``, ``data/pdb_all.tsv.gz`` and ``data/ec_all.tsv.gz``. It extracts the summary to ``data/go2name.tsv.gz``, ``data/is_a.tsv.gz`` and ``data/pdb_go.tsv.gz``.
This script reads swissprot name from ``uniprot/current_release/knowledgebase/complete/uniprot_sprot.fasta.gz`` and extract the summary to ``data/uniprot_sprot.tsv.gz``.

### Step 13: make foldseek database ###
This step must be performed after step 9
```bash
./script/make_foldseek.pl
```
Thi script creates foldseek database at ``foldseek/``

### Step 14: curate rhea annotation ###
This step must be performed after step 12.
```bash
./script/download_rhea.pl
```
This script downloads firedb, rhea and chebi to``firedb/``, ``rhea/``, and ``chebi``.
It then maps protein receptors to rhea according to ``rhea/uniprot2rhea.tsv``; if the uniprot protein is not in rhea, the protein is mapped to rhea according to its GO annotations at ``data/pdb_go.tsv.gz``.
The protein to rhea mappings are then written to ``data/pdb_rhea.tsv.gz``.
If a protein is present in this rhea mapping, the annotation score of the ligand interaction is calculated by the Tanimoto Coefficient between the ligand and the rhea ligands. Otherwise, the annotation score is categorized by firedb (cognate, ambiguous, non_coganate). The annotation score is saved to ``data/lig_rhea.tsv.gz``.

### Step 15: clean up intermediate files ###
This step must be run after everything is done.
```bash
./script/clean_up.pl
```

## Reference ##
Chengxin Zhang, Xi Zhang, Peter L Freddolino and Yang Zhang (2023)
[BioLiP2: an updated structure database for biologically relevant ligand-protein interactions](https://doi.org/10.1093/nar/gkad630).
Nucleic Acids Research, gkad630.

mmCIF2BioLiP

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README