263 step for alphafold multimer query json generation by AnnaPolensky · Pull Request #271 · cschlaffner/PROTzilla

AnnaPolensky · 2026-03-02T12:26:17Z

Description

fixes #263
Added a new Output-Tab "Downloads":

Therefore, introduced a new category of step methods, so that we have calc_method, plot_method and download_method.
Used the new download_method, to add a new step "AlphaFoldMultimerQueryJsonGeneration". User can input a list of uniprot ids and how many copies he wants to have of each protein. One can also add a seed if one wants to use a specific seed. Then a json-File is being generated that can be downloaded via the button in the Downloads-Tab.

Changes

backend/protzilla/importing/query_generation.py contains the method for generating the json
backend/protzilla/steps.py contains most of the changes for adding a download_method
frontend/src/components/app/run-screen/run-screen.tsx contains frontend changes for displaying the download-Tab and generating a button for each download.

Testing

Add the query generation step (section importing).
Enter protein ids (e.g. I tested with "O43242 O432432" and "2 2", although I believe that that is a query that does not make sense from a biological point of view).
Enter a seed or not. (Maybe try both versions.)
Calculate step.
Go to download-Tab and hit the download button.
See that the downloaded file follows this format: https://github.com/google-deepmind/alphafold/blob/main/server/README.md
(optional) Go to https://alphafoldserver.com/, continue with your Google-Account and check that you can upload the json:

Feel free to try different ids and different numbers of copies or try to break the step by putting in different input compared to the specified format.

PR checklist

Development

If necessary, I have updated the documentation (README, docstrings, etc.)
If necessary, I have created / updated tests.

Mergeability

main-branch has been merged into local branch to resolve conflicts
The tests and linter have passed AFTER local merge
The backend code has been formatted with black
The frontend code has been formatted with pnpm format and checked with pnpm lint

Code review

I have self-reviewed my code.
At least one other developer reviewed and approved the changes

…json-generation

…n were entered

github-actions · 2026-03-02T12:36:27Z

Coverage report

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
backend/main
views.py					624-650
views_settings.py					238-251, 255-259, 263-315, 326-340, 347-368, 372-447, 453, 464-482, 486-556, 562, 624-630
backend/protzilla
form.py
form_helper.py
networking.py					27-32
run.py					361, 365
steps.py					143-145, 245-252, 281, 308, 471, 474, 478, 483, 569-577, 681-689
backend/protzilla/data_analysis
crosslinking_validation.py					56, 246-247, 317-330
dimension_reduction.py
plots.py
backend/protzilla/data_integration
enrichment_analysis.py
enrichment_analysis_gsea.py
backend/protzilla/importing
alphafold_protein_structure_load.py					125, 208-211, 235-239, 304-307, 315-318, 361-362, 367, 371, 383, 416-420, 521-524, 556-562, 609-611, 654-656, 664-665, 674-682, 730-732, 743-744, 755-758, 863-864, 878, 900-902
crosslinking_import.py					148, 152, 202-205, 247-249, 254-259, 293-296, 311, 338-373, 395-429, 478, 480, 483-485, 684, 691, 693-696, 700, 710, 738-739, 774-778, 795-796, 801-802
import_utils.py
query_generation.py					57-64
backend/protzilla/methods
data_analysis.py					2558, 2616-2645
importing.py					440, 466, 502, 532, 598, 625
backend/protzilla/utilities
clustergram.py
utilities.py					142-144, 157-180
Project Total

_{This report was generated by python-coverage-comment-action}

…also comma-separated input

Elena-kal

One major problem is that inputting the uniprot ids like this: "P69905, P68871" does not work while "P69905 P68871" is fine. Should be fixed by stripping the whitespaces though.
Another issue I came across is that the step is validated (gets the green checkmark) even though there are error messages. This does not happen in other steps.
I also added a few suggestions to the code but overall the code seems fine.
I tested a few structures and used the json file for alphafold predictions. This worked very well. I am also very convinced of the downloads tab, I like it a lot and I think it could be useful for other steps as well.

Elena-kal · 2026-03-02T16:35:43Z

backend/protzilla/importing/query_generation.py

+
+    # extract protein_ids and number of copies per id and make sure they have the same length
+    if "," in protein_ids:
+        uniprot_ids = protein_ids.split(",")


Maybe also remove all whitespaces. I am not sure whether my problems came from that...

I think, too, that this caused your problem. I originally did not remove the whitespaces because I only added the comma-separated lists to support copy-paste from Excel where no whitespaces would be included. Changed this now.

Elena-kal · 2026-03-02T16:49:48Z

backend/protzilla/importing/query_generation.py

+        uniprot_ids = protein_ids.split()
+    try:
+        if "," in number_copies:
+            copies_per_id = [int(input) for input in number_copies.split(",")]


On the other hand, this split(",") worked I think

Because we are having an additional cast to an int here, so the cast kind of removes the whitespaces.

backend/protzilla/importing/query_generation.py

backend/protzilla/methods/importing.py

backend/tests/main/test_views_helper.py

Elena-kal · 2026-03-02T17:05:46Z

backend/protzilla/importing/query_generation.py

+        )
+    query_as_string = f"[{json.dumps(query)}]"
+    return dict(
+        messages={},


Why don't we return messages? And even if we want this to be empty, wouldn't we want the messages to be an empty list not a dict?

Thanks for pointing out, I added a success message.

backend/protzilla/importing/query_generation.py

Elena-kal · 2026-03-02T17:12:19Z

backend/protzilla/importing/query_generation.py

+    query_as_string = f"[{json.dumps(query)}]"
+    return dict(
+        messages={},
+        downloads={f"prediction_query_{'_'.join(uniprot_ids)}": query_as_string},


this file name could become very long if we use too many uniprot ids. maybe we could truncate it to prevent this.

I added another input field, the user now enters the filename themselves.

…ssage after successful query generation

AnnaPolensky · 2026-03-03T12:50:30Z

One major problem is that inputting the uniprot ids like this: "P69905, P68871" does not work while "P69905 P68871" is fine. Should be fixed by stripping the whitespaces though. Another issue I came across is that the step is validated (gets the green checkmark) even though there are error messages. This does not happen in other steps. I also added a few suggestions to the code but overall the code seems fine. I tested a few structures and used the json file for alphafold predictions. This worked very well. I am also very convinced of the downloads tab, I like it a lot and I think it could be useful for other steps as well.

You can now enter comma- and space-separated ids like "P69905, P68871". Also, changed the validation of the step. Happy to hear that you like the download tab :)

tE3m

some minor adjustments, but good changes overall

tE3m · 2026-03-03T15:50:57Z

backend/protzilla/methods/importing.py

+            input_fields=[
+                TextField(
+                    name="name",
+                    label="File name and AlphaFold job name for generated query",


this label should probably tell the user that the text they enter is only the stem of the filename, since .json is appended automatically

backend/main/views.py

backend/protzilla/importing/query_generation.py

…d-query-json-generation

AnnaPolensky added 8 commits February 16, 2026 11:04

feat: add step for generating a json alphafold multimer query

cb608b6

feat: add download tab to outputs

32c657f

merge: merge crosslinking into 263-step-for-alphafold-multimer-query-…

cfa88ab

…json-generation

fix: fix broken all_steps after merge

fe19dc6

feat: add download_methods

ce49260

feat: add download button for generated alphafold json queries

237f903

refactor: tidy up code

1268988

test: add tests for alphafold multimer query json generation

081a912

AnnaPolensky requested review from Elena-kal and tE3m March 2, 2026 12:26

AnnaPolensky self-assigned this Mar 2, 2026

feat: make sure that at least 2 protein ids or 2 copies of one protei…

0a817b5

…n were entered

AnnaPolensky added 2 commits March 2, 2026 15:19

feat: add input to use a specific seed and allow not only space- but …

d791ed8

…also comma-separated input

fix: fix broken test

068329e

Elena-kal requested changes Mar 2, 2026

View reviewed changes

AnnaPolensky added 3 commits March 3, 2026 13:17

fix: also accept input separated with comma and space, add success me…

ed7bb2a

…ssage after successful query generation

feat: add input field for file name of prediction query file

288e01a

fix: step only turns green if file was generated

2cdf9dd

tE3m reviewed Mar 3, 2026

View reviewed changes

AnnaPolensky added 3 commits March 3, 2026 17:58

fix: address code review feedback

c043e76

refactor: rename alphafold-multimer-query-json-generation to alphafol…

dedfba1

…d-query-json-generation

merge: merge crosslinking into 263

89e030b

Conversation

AnnaPolensky commented Mar 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Changes

Testing

PR checklist

Uh oh!

github-actions bot commented Mar 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Coverage report

Uh oh!

Elena-kal left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AnnaPolensky commented Mar 3, 2026

Uh oh!

tE3m left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

AnnaPolensky commented Mar 2, 2026 •

edited

Loading

github-actions bot commented Mar 2, 2026 •

edited

Loading