search

WayScience / phenotypic_profiling

Machine learning for predicting 15 single-cell phenotypes from cell morphology profiles
Creative Commons Attribution 4.0 International
1 stars 3 forks source link

[Response to Reviewers] Add Silhouette analysis #68

Closed gwaybio closed 5 days ago

gwaybio commented 6 days ago

This PR is in response to the following reviewer comment:

You state “By eye, CellProfiler features demonstrated the most heterogeneity…”, is there perhaps a way of quantifying this? For example, some kind of neighbourhood analysis.

We think this is a good idea, and therefore performed the following analysis:

  1. Per feature space, calculate Silhouette score per phenotype in an all vs. one comparison (e.g., All anaphase cells vs. all other cells)
  2. Apply PCA to make dimensions of input features consistent (n_components=50)
  3. Calculate average Silhouette width (per all-vs.-one phenotype and per feature space)

We interpret the Silhouette scores how well cells of a given phenotype are clustered compared to other cells of the same phenotype. A positive score means cells of the same phenotype are more similar to other cells of the same phenotype (on average) compared to all other cells. A score of 1 indicates complete separation of similar phenotypes from other phenotypes.

New supplementary figure

supplementary_silhouette_scores

review-notebook-app[bot] commented 6 days ago

Check out this pull request on  ReviewNB

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB

gwaybio commented 5 days ago

Thanks for the review @jenna-tomkinson - I caught a couple things too, which I addressed in the recent commits. Merging now!