The prankweb from cusbg

Check for initial status

It is possible that the inital qued status is shown even if the task is imediately started .. a solution would be to ignore the first response.

Add visualisation

Add cartoon + transparent surface visualisation.

The downloaded results archive contains a PDB file without any ligands

When the archive with the results is downloaded, it contains actually two PDB files, one in the root of the archive and one in the visualization/data folder. Both are named structure. However, the one, which is loaded with the pml script is actually a stripped version of the full PDB file and does not contain any information about any ligand.

When I try to run the predictions with the latest P2rank everything works well.

Update links to P2Rank

As P2Rank is not moving to cusbug we need to update links to it's current location.

Failed to predict 7bv2

10/22/2020 15:16:51 [INFO] - Preparing structure ...
10/22/2020 15:16:51 [DEBUG] - Downloading 'https://files.rcsb.org/download/7bv2.pdb' to '/data/p2rank/task/database/v2-conservation/7bv2_A,B,C,P,T/working/structure-raw.pdb' ...
10/22/2020 15:16:51 [DEBUG] - Starting new HTTPS connection (1): files.rcsb.org:443
10/22/2020 15:16:52 [DEBUG] - https://files.rcsb.org:443 "GET /download/7bv2.pdb HTTP/1.1" 200 167677
10/22/2020 15:16:52 [DEBUG] - Reading chains ...
/usr/local/lib/python3.7/dist-packages/Bio/PDB/StructureBuilder.py:92: PDBConstructionWarning: WARNING: Chain A is discontinuous at line 9455.
PDBConstructionWarning,
/usr/local/lib/python3.7/dist-packages/Bio/PDB/StructureBuilder.py:92: PDBConstructionWarning: WARNING: Chain P is discontinuous at line 9468.
PDBConstructionWarning,
/usr/local/lib/python3.7/dist-packages/Bio/PDB/StructureBuilder.py:92: PDBConstructionWarning: WARNING: Chain A is discontinuous at line 9492.
PDBConstructionWarning,
10/22/2020 15:16:52 [DEBUG] - Reading chains ... done
10/22/2020 15:16:52 [DEBUG] - Executing '/opt/protein-utils/bin/protein-utils -a filter-by-chain --structure /data/p2rank/task/database/v2-conservation/7bv2_A,B,C,P,T/working/structure-raw.pdb --output /data/p2rank/task/database/v2-conservation/7bv2_A,B,C,P,T/working/structure.pdb --chains T,C,P,A,B'
15:16:53 [main] INFO cz.siret.protein.utils.ApplicationEntry - Using action: filter-by-chain
15:16:53 [main] WARN o.b.nbio.structure.align.util.UserConfiguration - Could not read dir from system property PDB_DIR or environment variable PDB_DIR, using system's temp directory /tmp
15:16:53 [main] INFO o.b.nbio.structure.align.util.UserConfiguration - Could not read cache dir from system property PDB_CACHE_DIR or environment variable PDB_CACHE_DIR, using PDB directory instead /tmp/
15:16:53 [main] INFO org.biojava.nbio.structure.io.PDBFileParser - Could not parse revision date string ''.
15:16:53 [main] INFO org.biojava.nbio.structure.io.PDBFileParser - Could not parse revision date string ''.
15:16:53 [main] INFO org.biojava.nbio.structure.io.PDBFileParser - Could not parse revision date string ''.
WARNING: An illegal reflective access operation has occurred
WARNING: Illegal reflective access by com.sun.xml.bind.v2.runtime.reflect.opt.Injector (file:/opt/protein-utils/lib/jaxb-impl-2.3.0.jar) to method java.lang.ClassLoader.defineClass(java.lang.String,byte[],int,int)
WARNING: Please consider reporting this to the maintainers of com.sun.xml.bind.v2.runtime.reflect.opt.Injector
WARNING: Use --illegal-access=warn to enable warnings of further illegal reflective access operations
WARNING: All illegal access operations will be denied in a future release
15:16:55 [main] INFO cz.siret.protein.utils.ApplicationEntry - Finished in 00:00:02
10/22/2020 15:16:55 [INFO] - Preparing structure ... done
10/22/2020 15:16:55 [DEBUG] - Path: /data/p2rank/task/database/v2-conservation/7bv2_A,B,C,P,T/working/structure.pdb
10/22/2020 15:16:55 [DEBUG] - Chains: T C P A B
10/22/2020 15:16:55 [DEBUG] - Executing '/opt/protein-utils/bin/protein-utils -a extract-chain-sequence --structure /data/p2rank/task/database/v2-conservation/7bv2_A,B,C,P,T/working/structure.pdb --output /data/p2rank/task/database/v2-conservation/7bv2_A,B,C,P,T/working/conservation-T/sequence.fasta --chain T'
15:16:55 [main] INFO cz.siret.protein.utils.ApplicationEntry - Using action: extract-chain-sequence
15:16:55 [main] WARN o.b.nbio.structure.align.util.UserConfiguration - Could not read dir from system property PDB_DIR or environment variable PDB_DIR, using system's temp directory /tmp
15:16:55 [main] INFO o.b.nbio.structure.align.util.UserConfiguration - Could not read cache dir from system property PDB_CACHE_DIR or environment variable PDB_CACHE_DIR, using PDB directory instead /tmp/
15:16:55 [main] INFO org.biojava.nbio.structure.io.PDBFileParser - Entity information (COMPOUND record) not found in file. Will assign entities heuristically
15:16:56 [main] ERROR cz.siret.protein.utils.ApplicationEntry - Action execution failed
15:16:56 [main] INFO cz.siret.protein.utils.ApplicationEntry - Reason
cz.siret.protein.utils.action.ActionFailed: Missing chain: T
at cz.siret.protein.utils.action.ChainToSequence.getChainSequence(ChainToSequence.java:19)
at cz.siret.protein.utils.command.StructureToFasta.saveSequence(StructureToFasta.java:78)
at cz.siret.protein.utils.command.StructureToFasta.execute(StructureToFasta.java:64)
at cz.siret.protein.utils.command.StructureToFasta.execute(StructureToFasta.java:20)
at cz.siret.protein.utils.ApplicationEntry.executeAction(ApplicationEntry.java:243)
at cz.siret.protein.utils.ApplicationEntry.executeAction(ApplicationEntry.java:204)
at cz.siret.protein.utils.ApplicationEntry.run(ApplicationEntry.java:55)
at cz.siret.protein.utils.ApplicationEntry.main(ApplicationEntry.java:41)
15:16:56 [main] INFO cz.siret.protein.utils.ApplicationEntry - Finished in 00:00:00
Traceback (most recent call last):
File "/opt/p2rank-runtime/run_p2rank_task.py", line 359, in <module>
main(_read_arguments())
File "/opt/p2rank-runtime/run_p2rank_task.py", line 64, in main
structure_file, chains, configuration, arguments)
File "/opt/p2rank-runtime/run_p2rank_task.py", line 190, in prepare_conservation
return compute_from_structure(structure_file, chains, arguments)
File "/opt/p2rank-runtime/run_p2rank_task.py", line 236, in compute_from_structure
for chain in chains
File "/opt/p2rank-runtime/run_p2rank_task.py", line 236, in <dictcomp>
for chain in chains
File "/opt/p2rank-runtime/run_p2rank_task.py", line 255, in compute_from_structure_for_chain
conservation.compute_conservation(fasta_file, working_dir, target_file)
File "/opt/p2rank-runtime/conservation.py", line 57, in compute_conservation
compute_msa(input_file, working_dir, msa_file)
File "/opt/p2rank-runtime/conservation.py", line 64, in compute_msa
sequences = _read_fasta_file(fasta_file)
File "/opt/p2rank-runtime/conservation.py", line 84, in _read_fasta_file
with open(input_file) as in_stream:
FileNotFoundError: [Errno 2] No such file or directory: '/data/p2rank/task/database/v2-conservation/7bv2_A,B,C,P,T/working/conservation-T/sequence.fasta'

Add uniprot30 as a BLAST database

As searching uniprot90 takes a long time we should try to use uniprot30 as another step to boost the performance.

Hyde AA when binding site is deselected

Now only the colour disappear, but in some visualisation, the AA are still visible.

Fix protein-utils warning

From log

WARNING: An illegal reflective access operation has occurred
WARNING: Illegal reflective access by com.sun.xml.bind.v2.runtime.reflect.opt.Injector (file:/opt/protein-utils/lib/jaxb-impl-2.3.0.jar) to method java.lang.ClassLoader.defineClass(java.lang.String,byte[],int,int)
WARNING: Please consider reporting this to the maintainers of com.sun.xml.bind.v2.runtime.reflect.opt.Injector
WARNING: Use --illegal-access=warn to enable warnings of further illegal reflective access operations
WARNING: All illegal access operations will be denied in a future release

This is issue with jaxb-impl 2.3.0 use by BioJava.
The jaxb-impl 2.4.0 should be fixed, but it is not used by BioJava.
So we need to monitor BioJava for updates.

Allow standalone execution of p2rank runtime docker

Add a tutorial, and describe how to run with/without conservartion and with CSV feature (custom branch).

Add citation

Add citation https://doi.org/10.1093/nar/gkz424 into index footer, and view footer.
Add information into about page with cite for above and p2rank.

Check Docker files with hadolint

https://hadolint.github.io/hadolint/

Update custom MSA input

We need to allow user to upload multiple MSA files, one for each chain.
Format as produced by https://www.ebi.ac.uk/Tools/msa/muscle/

Missing binding site information

No pockets found

Add explicit message with information that no pockets were found. As a bonus we can offer to re-run the prediction using more chains.

Example is 7BV2 for chain B. As of now it may not be clear to the user that there are no pockets.

Add maintenance page

We should replace default 502 Bad Gateway site with something better. This apply to the server configuration not the project it self.

We should also have some sort of message if any of the project components are offline.

https://serverfault.com/questions/426347/nginx-reverse-proxy-error-page

The log should not be red

Assign different colors to the debug, info, and warning messages which appear when waiting for the results.

Use of HTTPS resources

Even if the web is running on HTTPS it is fetching (jQuery UI, ...) resources from HTTP, this make the site "unsafe". In order to fix this we need to update the dependencies .

Protael - http://sanshu.github.io/protaelweb/getting-started.html

Re-use conservation scores inside a structure

For some structures, i.e. homomultimers, there is a same chain multiple times. We need to be able to detect this and do not run conservation on the same sequence multiple times.

As in future the conservation pipeline should be able to cache data of computed sequences, the implemented solution should be removable in the future.

Add deployment specific docker-compose

Now we have a docker-compose.production file, this is not enought as we have multiple environments. There should be a file for each env. and also for running on a local machine.

This shold include configuration of env properties such as number of workers and GA tags.

ELIXIR integration

On the welcome page and on the result page there is Elixir logo with no links or explanation. What is Elixir and what else it can do for me?

How it may look like: https://webchem.ncbr.muni.cz/Platform/ChargeCalculator

Also, P2Rank/PrankWeb is not listed among Elixir services https://www.elixir-czech.cz/services

Improve handling of inputs with no protein atoms

For example 7BV2 chain P.

A solution is to add a number of protein atoms into output of protein-utils and check this value in the Python script.

Remove wiki

Remove wiki or redirect it to https://github.com/cusbg/p2rank-framework/wiki

Validate nginx configuration

We can use https://github.com/yandex/gixy

Show loading indicator as soon as possible

Update to P2Rank 2.2

With new version there are following relevant changes:

Add time stamps to log -stdout_timestamp "yyyy.MM.dd HHmm:"
Update conservation file name to {basename}{chain}.(.*).hom(.gz|.zst|.bz2|{})
where {basename} is one of
{basename}.pdb
{basename}.pdb.gz
{basename}.ent
{basename}.ent.gz
{basename}.cif
{basename}.cif.gz

Can't load MSA (or any file) from a file

On Mac

Show chains after return from the results page

When the user returns from the results page to the main page, the structure chains do not show. To make them appear, one has to reenter the PDB ID.

Document ability to run batch predictions via API

Add comment to the "PDB Code"

This can be selected only when user provide a custom strcuture file.
Perhaps we can display this field only after user select strcuture file.

Add logo to NavBar

Replace project name with project logo.

Update help

Move and update help page to reflect changes in UI. The help page should be part of the repository Wiki.

Update deployment

Prepare deployment using:

Implement GA component

Is your feature request related to a problem? Please describe.
We need to be able to monitor GET requests made to a certain endpoint using Google Analytics.

Describe the solution you'd like
Add optional NodeJs backend component (Docker) with redirect from api-gateway.

Describe alternatives you've considered

Additional context

https://developers.google.com/analytics/devguides/collection/protocol/v1/devguide

Bundle js libraries to the web app

Revise use of nginx

https://hub.docker.com/_/nginx

Entrypoint quiet logs
Running nginx as a non-root user
Amplify
https://www.nginx.com/resources/wiki/start/topics/tutorials/config_pitfalls/

Rename the file to be downloaded

Use the PDB code or file name (in case the user uploaded a file)

Test support for custom MSA

We need to support https://www.ebi.ac.uk/Tools/msa/muscle/

No stripping of white space from PDB codes

When the user pastes a PDB code with trailing spaces, it is not recognized. The correct behavior would be to always strip the white spaces. The problem is that only without the trailing spaces, the code is recognized and chain selection is offered.

Add link to file a github issue on the web

on the main page or about pae

Update sequence filtering

As of now, we limit the number of sequences given to MUSCLE by just using top-n. In order to increase diversity we should choose each m-th sequence instead.

Add deployment automation

We should redeploy from master branch.

https://docs.github.com/en/free-pro-team@latest/actions/guides/publishing-docker-images

Publication link for PrankWeb article is missing

on site http://prankweb.cz/views/download.jsp

'Show only this binding site' button

Now there is a button to focus a camera on the binding site, we need to add another button that will hyde all other binding sites.

Leave original PDB file for visualisation

As of now with example 2SRC there is no bindings sites / ligands visible in the visualisation. They should be.

Submitting an empty protein

When the user deselects all chains at the input, the computation proceeds (starts to collect MSA and so on) and even does the prediction. Instead, we should maybe disable the submit button and, in the optimal case, show a warning/error message.

Performance issues with protein utils

Describe the bug
Running protein-utils on 19HC, cause performance issues when clustering ligands.

To Reproduce

Run .\protein-utils\bin\protein-utils.bat -a p2rank-web --structure=./19HC.pdb --prediction=./19hc.pdb.gz_predictions.csv --residues=./19hc.pdb.gz_residues.csv --output-pocket=./prediction.json --output-sequence=./sequence.json

Expected behavior
The execution finish in reasonable time.

09/09/2020 07:37:55 [DEBUG] - Executing 'cd /opt/conservation/jense_shannon_divergence/ && 
python2 score_conservation.py /data/p2rank/task/database/v2-conservation/2SRC/working/conservation-A/msa > /data/p2rank/task/database/v2-conservation/2SRC/working/conservation-A/structure_A.score'
Traceback (most recent call last):
File "score_conservation.py", line 831, in <module>
seq_weights = load_sequence_weights(align_file.replace('.%s' % align_suffix, '.weights'))
File "score_conservation.py", line 599, in load_sequence_weights
seq_weights.append(float(l[1]))
ValueError: could not convert string to float: A
Traceback (most recent call last):
File "/opt/p2rank-runtime/execute_p2rank.py", line 359, in <module>
main(_read_arguments())
File "/opt/p2rank-runtime/execute_p2rank.py", line 65, in main
structure_file, chains, configuration, arguments)
File "/opt/p2rank-runtime/execute_p2rank.py", line 190, in prepare_conservation
return compute_from_structure(structure_file, chains, arguments)
File "/opt/p2rank-runtime/execute_p2rank.py", line 236, in compute_from_structure
for chain in chains
File "/opt/p2rank-runtime/execute_p2rank.py", line 236, in <dictcomp>
for chain in chains
File "/opt/p2rank-runtime/execute_p2rank.py", line 255, in compute_from_structure_for_chain
conservation.compute_conservation(fasta_file, working_dir, target_file)
File "/opt/p2rank-runtime/conservation.py", line 58, in compute_conservation
compute_jensen_shannon_divergence(msa_file, output_file)
File "/opt/p2rank-runtime/conservation.py", line 257, in compute_jensen_shannon_divergence
execute_command(cmd)
File "/opt/p2rank-runtime/execute_p2rank.py", line 170, in execute_command
result.check_returncode()
File "/usr/lib/python3.7/subprocess.py", line 428, in check_returncode
self.stderr)
subprocess.CalledProcessError: Command 'cd /opt/conservation/jense_shannon_divergence/ && python2 score_conservation.py /data/p2rank/task/database/v2-conservation/2SRC/working/conservation-A/msa > /data/p2rank/task/database/v2-conservation/2SRC/working/conservation-A/structure_A.score' returned non-zero exit status 1.

cusbg / prankweb Goto Github PK

prankweb's People

Contributors

Stargazers

Watchers

Forkers

prankweb's Issues

Recommend Projects

Recommend Topics

Recommend Org