Skip to content
Snippets Groups Projects
Commit e8a47274 authored by Bienchen's avatar Bienchen
Browse files

Proofing README.md

parent b7926581
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
Showing
with 97 additions and 72 deletions
docker/README.md
+ 97
− 72
View file @ e8a47274
......@@ -6,23 +6,47 @@ https://swissmodel.expasy.org/qmean/),
[Git](https://git.scicore.unibas.ch/schwede/QMEAN),
[Docker](https://git.scicore.unibas.ch/schwede/QMEAN/container_registry)).
Table Of Contents
-----------------
* [Available Scoring Functions](#scoringfunctions)
* [Input Requirements](#inputrequirements)
* [Structure Processing](#structureprocessing)
* [Obtain the image (Docker `pull`)](#qmeanpull)
* [Additional requirements](#additionalrequirements)
* [Score](#score)
* [Singularity](#singularity)
* [Results](#results)
* [Examples](#examples)
<a name="scoringfunctions"></a>Available Scoring Functions
----------------------------------------------------------
The following scoring functions are implemented:
[QMEANDisCo](https://doi.org/10.1093/bioinformatics/btz828):
> Studer, G., Rempfer, C., Waterhouse, A.M., Gumienny, R., Haas, J., Schwede, T. QMEANDisCo-distance constraints applied on model quality estimation, Bioinformatics 36, 1765–1771 (2020).
>
[QMEAN](https://doi.org/10.1093/bioinformatics/btq662)
[QMEAN](https://doi.org/10.1093/bioinformatics/btq662):
> Benkert, P., Biasini, M., Schwede, T. Toward the estimation of the absolute quality of individual protein structure models. Bioinformatics 27, 343-350 (2011).
>
[QMEANBrane](https://doi.org/10.1093/bioinformatics/btu457)
[QMEANBrane](https://doi.org/10.1093/bioinformatics/btu457):
> Studer, G., Biasini, M., Schwede, T. Assessing the local structural quality of transmembrane protein models using statistical potentials (QMEANBrane), Bioinformatics 30, i505–i511 (2014).
>
For a short description of the different scoring functions refer to the QMEAN
server [help page](https://swissmodel.expasy.org/qmean/help).
Input Requirements
------------------
<a name="inputrequirements"></a>Input Requirements
--------------------------------------------------
The container can read protein structures in PDB (.pdb) or mmCIF (.cif/.mmcif)
format. Compressed files are accepted, too (.gz suffix, e.g. 1crn.pdb.gz).
......@@ -42,20 +66,20 @@ in a single FASTA file.
The container calculates sequence profiles using
HHblits ([DOI](https://doi.org/10.1186/s12859-019-3019-7),
[Git](https://github.com/soedinglab/hh-suite)) for each unique SEQRES sequence.
If you already have the respective profiles available in a3m format, you can
If you already have the respective profiles available in A3M format, you can
speed things up (via the option `--profiles`). This only works if you also
provide SEQRES as an input and the master sequence for each profile must match
one of the SEQRES sequences.
Structure Processing
--------------------
<a name="structureprocessing"></a>Structure Processing
------------------------------------------------------
Structures are processed with the MOLecular ChecKer
Structures are processed with the **MOL**ecule **C**hec**K**er
([molck](https://openstructure.org/docs/dev/mol/alg/molalg/?highlight=molck#ost.mol.alg.Molck))
before scoring. In detail:
* Non-standard residues are mapped to their standard counterparts if possible
(e.g. Phospho-Tyrosine to Tyrosine or Seleno-Methionine to Methionine etc.).
(e.g. Phospho-Tyrosine to Tyrosine, Seleno-Methionine to Methionine, etc.).
Mapping information is derived from the
[component dictionary](http://www.wwpdb.org/data/ccd)
* Hydrogen atoms are stripped
......@@ -63,7 +87,7 @@ before scoring. In detail:
* Everything except the 20 standard proteinogenic amino acids is stripped
(after potential mapping of non-standard residues above)
* Unknown atoms are stripped, i.e. atoms that are not expected based on the
[component dictionary](http://www.wwpdb.org/data/ccd).
[component dictionary](http://www.wwpdb.org/data/ccd), as found in some force fields.
* Chains are potentially renamed, i.e. if a chain name is ' ', it gets
assigned a valid chain name.
......@@ -87,21 +111,21 @@ registry.scicore.unibas.ch/schwede/qmean:4.2.0
$
```
Additional requirements
-----------------------
<a name="additionalrequirements"></a>Additional requirements
------------------------------------------------------------
We need the non-redundant
[UniClust30 sequence database](https://uniclust.mmseqs.com/) to build sequence
profiles with HHblits. The following files are required:
* X_a3m.ffdata
* X_a3m.ffindex
* X_hhm.ffdata
* X_hhm.ffindex
* X_cs219.ffdata
* X_cs219.ffindex
* `X_a3m.ffdata`
* `X_a3m.ffindex`
* `X_hhm.ffdata`
* `X_hhm.ffindex`
* `X_cs219.ffdata`
* `X_cs219.ffindex`
with X being your UniClust30 version of choice. The productive QMEAN server uses
with `X` being your UniClust30 version of choice. The productive QMEAN server uses
UniClust30 [August 2018](http://wwwuser.gwdg.de/~compbiol/uniclust/2018_08/).
The directory with the files must be mounted to the container:
......@@ -115,16 +139,16 @@ https://swissmodel.expasy.org/repository/download/qmtl/qmtl.tar.bz2). The QMTL
is updated weekly following the PDB update cycle. Checking for new versions on
a Thursday is a good idea. The following files are required:
* smtl_uniq_cs219.ffdata
* smtl_uniq_cs219.ffindex
* smtl_uniq_hhm.ffdata
* smtl_uniq_hhm.ffindex
* CHAINCLUSTERINDEX
* indexer.dat
* seqres_data.dat
* atomseq_data.dat
* ca_pos_data.dat
* VERSION
* `smtl_uniq_cs219.ffdata`
* `smtl_uniq_cs219.ffindex`
* `smtl_uniq_hhm.ffdata`
* `smtl_uniq_hhm.ffindex`
* `CHAINCLUSTERINDEX`
* `indexer.dat`
* `seqres_data.dat`
* `atomseq_data.dat`
* `ca_pos_data.dat`
* `VERSION`
Again, the corresponding directory must be mounted:
......@@ -132,11 +156,11 @@ Again, the corresponding directory must be mounted:
-v <PATH_TO_LOCAL_QMTL>:/qmtl
```
Score
-----
<a name="score"></a>Score
-------------------------
Having everything setup, you can score model.pdb with SEQRES data stored in
seqres.fasta using QMEANDisCo:
Having everything setup, you can score `model.pdb` with SEQRES data stored in
`seqres.fasta` using QMEANDisCo:
```terminal
docker run --workdir $(pwd) -v $(pwd):$(pwd) -v <PATH_TO_LOCAL_UNICLUST>:/uniclust30 -v <PATH_TO_LOCAL_QMTL>:/qmtl registry.scicore.unibas.ch/schwede/qmean:4.2.0 run_qmean.py model.pdb --seqres seqres.fasta
......@@ -152,71 +176,72 @@ The following gives more details on additional command line arguments:
docker run registry.scicore.unibas.ch/schwede/qmean:4.2.0 run_qmean.py --help
```
Singularity
-----------
<a name="singularity"></a>Singularity
-------------------------------------
A Singularity Image can directly be pulled from our registry:
A Singularity Image can directly be pulled & build from our registry:
```terminal
singularity build qmean_container.sif docker://registry.scicore.unibas.ch/schwede/qmean:4.2.0
```
Singularity directly allows to access the current working directory from within the container
Singularity allows to directly access the current working directory from within the container,
so scoring simplifies to:
```terminal
singularity run -B <PATH_TO_LOCAL_UNICLUST>:/uniclust30 -B <PATH_TO_LOCAL_QMTL>:/qmtl qmean_container.sif run_qmean.py model.pdb --seqres seqres.fasta
```
Results
-------
<a name="results"></a>Results
-----------------------------
Results are json formatted. For each model there is an entry with following
keys:
* chains: Contains the ATOMSEQ/SEQRES mapping for each chain. The ATOMSEQ is
* `chains`: Contains the ATOMSEQ/SEQRES mapping for each chain. The ATOMSEQ is
extracted from the structure, SEQRES is provided by the user. If SEQRES is not
provided => SEQRES == ATOMSEQ
* original_name: Filename of the input model
* preprocessing: Summarizes the outcome of input structure processing described
provided, SEQRES equals the ATOMSEQ
* `original_name`: Filename of the input model
* `preprocessing`: Summarizes the outcome of input structure processing described
above
* scores: Model specific scores. No matter what scoring method you're running
* `scores`: Model specific scores. No matter what scoring method you're running
(QMEAN [1] /QMEANDisCo [2] /QMEANBrane [3]), you have two keys:
"local_scores" and "global_scores". While the data in "global_scores"
calculates values according [1], the "local_scores" are method dependent.
For "global_scores" you get another dictionary with keys:
* "acc_agreement_norm_score"
* "acc_agreement_z_score"
* "avg_local_score" (mode dependent)
* "avg_local_score_error" (only set for [2])
* "cbeta_norm_score"
* "cbeta_z_score"
* "interaction_norm_score"
* "interaction_z_score"
* "packing_norm_score"
* "packing_z_score"
* "qmean4_norm_score"
* "qmean4_z_score"
* "qmean6_norm_score"
* "qmean6_z_score"
* "ss_agreement_norm_score"
* "ss_agreement_z_score"
* "torsion_norm_score"
* "torsion_z_score"
For "local_scores" you get another dictionary with chain names as keys and
`local_scores` and `global_scores`. While the data in `global_scores`
calculates values according [1], the `local_scores` are method dependent.
For `global_scores` you get another dictionary with keys:
* `acc_agreement_norm_score`
* `acc_agreement_z_score`
* `avg_local_score` (mode dependent)
* `avg_local_score_error` (only set for [2])
* `cbeta_norm_score`
* `cbeta_z_score`
* `nteraction_norm_score`
* `interaction_z_score`
* `packing_norm_score`
* `packing_z_score`
* `qmean4_norm_score`
* `qmean4_z_score`
* `qmean6_norm_score`
* `qmean6_z_score`
* `ss_agreement_norm_score`
* `ss_agreement_z_score`
* `torsion_norm_score`
* `torsion_z_score`
For `local_scores` you get another dictionary with chain names as keys and
lists with local scores as value. The local score lists have the same length
as the according SEQRES (ATOMSEQ if SEQRES is not given as an input).
as the corresponding SEQRES (ATOMSEQ if SEQRES is not given as an input).
The location of a residue in SEQRES determines the location of its local score
in the score list.
Examples
--------
<a name="examples"></a>Examples
-------------------------------
Example data to run all available scoring functions are available in the
qmean_qmeandisco_example and qmeanbrane_example directory.
[qmean_qmeandisco_example](docker/qmean_qmeandisco_example) and
[qmeanbrane_example](docker/qmeanbrane_example) directory.
[comment]: <> ( LocalWords: QMEANDisCo mmCIF JSON GitLab DBeacons cd OST )
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment