Add function for feature prioritization by AbhirupaGhosh · Pull Request #17 · JRaviLab/amRml

AbhirupaGhosh · 2026-03-13T19:52:53Z

This function computes feature improvement by reading and processing feature data from parquet files, applying various transformations and calculations to derive insights on feature importance and contributions.

Description

What kind of change(s) are included?

Feature (adds or updates new capabilities)
Bug fix (fixes an issue).
Enhancement (adds functionality).
Breaking change (these changes would cause existing functionality to not work as expected).

Checklist

Please ensure that all boxes are checked before indicating that this pull request is ready for review.

I have read and followed the CONTRIBUTING.md guidelines.
I have searched for existing content to ensure this is not a duplicate.
I have performed a self-review of these additions (including spelling, grammar, and related).
I have added comments to my code to help provide understanding.
I have added a test which covers the code changes found within this PR.
I have deleted all non-relevant text in this pull request template.
Reviewer assignment: Tag a relevant team member to review and approve the changes.

This function computes feature improvement by reading and processing feature data from parquet files, applying various transformations and calculations to derive insights on feature importance and contributions.

Added computeFeatureScore function to calculate feature scores based on various metrics and cluster information.

…ion and introducing wide table functions for global score calculations

…update parameters for clarity

…improve clarity

… feature and cluster analysis

…mprove clarity

…b/amRml into feature_rescoring_simple

AbhirupaGhosh · 2026-07-24T18:58:07Z

This script introduces a workflow for prioritizing molecular features associated with resistance across drugs and drug classes.

This workflow identifies features that are consistently important across repeated model fits (seeds), making the results less noisy and more interpretable.

The workflow:

scores and summarizes feature importance across seeds,
prioritizes stable, high-ranking features for each drug or drug class,
aggregates prioritized features into protein clusters,
identifies shared and unique clusters across drugs/classes, and
builds feature–cluster networks for visualization and downstream interpretation.

The run would be

all_top_features_parquet <- "inst/extdata/all_top_features.parquet"
scored_features <- scoreFeaturesWithinSeed(all_top_features_parquet)
feature_summary <- summariseFeaturesAcrossSeeds(scored_features)

top_features <- topFeaturesPerDrugOrClass(feature_summary,
                                        rank_score_quantile = 0.95,
                                        threshold_sd_rank = 1)

cluster_feature_parquet <- "inst/extdata/cluster_feature.parquet"
top_clusters <- summariseClusters(top_features, cluster_feature_parquet)
protein_names_parquet <- "inst/extdata/protein_names.parquet"

feature_network <- buildFeatureNetwork(top_features, top_clusters, 
cluster_feature_parquet, protein_names_parquet)

plotFeatureNetworkD3(feature_network)

jananiravi · 2026-07-28T19:32:42Z

Minor suggestions for now or later:

in example runs, changed = to <-
species = full_name, spp = 3_letter_name?
there's a mix of small and caps (which I'm editing)

Sign gets fixed in the next one feature_summary but Variable is still in caps.

@AbhirupaGhosh, can you clarify the following:

Variable = feature_name?
what's 'signal network'?

Standardize feature column names across the codebase (Variable -> variable, Importance -> importance, Sign -> sign) and update all uses (mutate, group_by, joins, transmute, distinct). Rename summariseFeatureAcrossSeeds -> summariseFeaturesAcrossSeeds, buildSignalNetwork -> buildFeatureNetwork, and plotSignalNetworkD3 -> plotFeatureNetworkD3. Adjust join keys (variable == feature), grouping and summarise calls, examples, and roxygen docs to match the new names. Includes minor formatting and whitespace cleanups. These changes unify naming and clarify the feature/cluster network API.

jananiravi

Added a few minor fixes inline + some comments for clarification.

All other scripts are named do_something. so 'prioritize_features' or 'rescore_features' or something like that?

Note from Slack
In summariseClusters() and buildFeatureNetwork(), when a top-ranked domain/COG feature gets joined to cluster_feature, it fans out into every cluster carrying that domain — including clusters with no real connection to the resistance signal, just shared domain architecture.

Concretely:

A single promiscuous COG (fan-out up to 189) inflates frequency/n_variables/cluster_mean_rank_score for up to 189 different clusters in summariseClusters(), none of which may be the true driver.
In buildFeatureNetwork(), feature_cluster_edges will make that one feature a hub node connected to up to ~189 cluster nodes, which will dominate/clutter plotFeatureNetworkD3()'s output and overstate specificity.

So, is this fan-out intentional (i.e., "show all plausible clusters a domain could implicate"), or should there be a cap/weighting (e.g., down-weight edges by 1/n_clusters, or exclude domains above some fan-out threshold) so ubiquitous COGs/Pfams don't swamp the network? This is a modeling decision, not a coding bug.

jananiravi · 2026-07-28T20:01:56Z

-  protein_scale_realization <- ranked_features |>
-    dplyr::filter(!is.na(cluster), shuffled == FALSE) |>
+  top_clusters <- top_features |>
+    dplyr::left_join(cluster_feature, by = dplyr::join_by(Variable == feature)) |>


⚠️ Warning message:

In dplyr::left_join(top_features, cluster_feature, by = dplyr::join_by(variable == : Detected an unexpected many-to-many relationship between `x` and `y`. ℹ Row 1 of `x` matches multiple rows in `y`. ℹ Row 54683 of `y` matches multiple rows in `x`. ℹ If a many-to-many relationship is expected, set `relationship = "many-to-many"` to silence this warning.

jananiravi · 2026-07-28T20:03:18Z

+
+  cluster_table <- top_clusters |>
+    dplyr::mutate(model_id = make_model_id(drug_label, drug_or_class)) |>
+    dplyr::left_join(


⚠️ Warning message:

In dplyr::left_join(feature_table, cluster_feature, by = dplyr::join_by(variable == : Detected an unexpected many-to-many relationship between `x` and `y`. ℹ Row 1 of `x` matches multiple rows in `y`. ℹ Row 151601 of `y` matches multiple rows in `x`. ℹ If a many-to-many relationship is expected, set `relationship = "many-to-many"` to silence this warning.

jananiravi · 2026-07-28T20:26:54Z

+  dplyr::ungroup() |>
+  dplyr::group_by(drug_label, drug_or_class) |>
+  dplyr::filter(
+  #   n_seeds == max(n_seeds),


intentionally left out? what's the consistency across seeds you're looking for?

yes, I am using the threshold_sd_rank as a proxy for consistency.
max(n_seeds) is too stringent

jananiravi · 2026-07-28T20:28:16Z

    ) |>
    dplyr::mutate(
-      contribution = Importance / sum(Importance, na.rm = TRUE),
+      contribution = importance / sum(importance, na.rm = TRUE),


cumulative impact first then rank?

ranking based on the contribution or importance will give the same. I kept contribution if we want to care about contribution %tile.

jananiravi · 2026-07-28T20:28:53Z

+    mean_rank_score = mean(rank_score, na.rm = TRUE),
+    best_rank = min(rank, na.rm = TRUE),
+    rank_consistent = dplyr::n_distinct(rank) == 1,
+    rank_sd = sd(rank, na.rm = TRUE),


rank_score, mean_rank_score are normalized but not this one? are these comparable?

rank_score_sd instead?

I used rank_sd because it is easy to dry run with whole numbers.

jananiravi · 2026-07-28T22:23:03Z

-      contribution = Importance / sum(Importance, na.rm = TRUE),
+      contribution = importance / sum(importance, na.rm = TRUE),
      rank = dplyr::dense_rank(dplyr::desc(contribution)),
      n_features = dplyr::n(),


Is this to get totals by feature scale? If so, group_by then n() within summarize?

Add computeFeatureImprovement function for feature analysis

8d6f2f4

This function computes feature improvement by reading and processing feature data from parquet files, applying various transformations and calculations to derive insights on feature importance and contributions.

AbhirupaGhosh requested review from charmvang, epbrenner and jananiravi March 13, 2026 19:52

AbhirupaGhosh self-assigned this Mar 13, 2026

AbhirupaGhosh and others added 3 commits March 13, 2026 19:57

Style code (GHA)

729c837

Implement computeFeatureScore function for feature scoring

62668b4

Added computeFeatureScore function to calculate feature scores based on various metrics and cluster information.

Style code (GHA)

a388841

AbhirupaGhosh marked this pull request as draft April 10, 2026 21:25

AbhirupaGhosh added 8 commits July 22, 2026 10:38

Update feature_rescoring.R

4098ed6

Update feature_rescoring.R

82964cc

Refactor feature scoring functions and enhance documentation for clarity

b356109

Enhance feature rescoring by adding distinct feature subtype aggregat…

b96d438

…ion and introducing wide table functions for global score calculations

Refactor buildClusterWideTable function to improve documentation and …

ba1faf2

…update parameters for clarity

Refactor buildClusterWideTable function to streamline parameters and …

0b617d3

…improve clarity

Implement signal network construction and visualization functions for…

5ac44a4

… feature and cluster analysis

Refactor buildSignalNetwork function to remove unused parameter and i…

e29a80f

…mprove clarity

AbhirupaGhosh changed the title ~~Add computeFeatureImprovement function for feature analysis~~ Add function for feature prioritization Jul 24, 2026

AbhirupaGhosh added 3 commits July 24, 2026 12:45

Merge branch 'main' into feature_rescoring_simple

b715216

Remove unused functions

708543a

Merge branch 'feature_rescoring_simple' of https://github.com/JRaviLa…

4b9d555

…b/amRml into feature_rescoring_simple

AbhirupaGhosh marked this pull request as ready for review July 24, 2026 18:57

AbhirupaGhosh requested review from amcim and eboyer221 July 24, 2026 18:58

AbhirupaGhosh added 2 commits July 28, 2026 10:54

Merge branch 'main' into feature_rescoring_simple

9723262

Add example data and remove old, unused data

b8f45b1

jananiravi requested changes Jul 28, 2026

View reviewed changes

Update feature_rescoring.R

b21796f

jananiravi reviewed Jul 28, 2026

View reviewed changes

Uh oh!

Conversation

AbhirupaGhosh commented Mar 13, 2026

Description

What kind of change(s) are included?

Checklist

Uh oh!

AbhirupaGhosh commented Jul 24, 2026 • edited by jananiravi Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jananiravi commented Jul 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jananiravi 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!

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!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

AbhirupaGhosh commented Jul 24, 2026 •

edited by jananiravi

Loading

jananiravi commented Jul 28, 2026 •

edited

Loading