[Response to Reviewers] Update dimensionality reduction figures

[gwaybio]

This PR accomplishes two, interrelated items, that I perform in response to two separate reviewer comments.

1) Move supplementary UMAP to main figure 2

Reviewer comment

Fig 2A – From this figure it looks like “elongated”, “large” and “interphase” (rather than “metaphase”) appear to be most distinct for CellProfiler features; however, looking at the individual maps in Fig S2, the split of “interphase” is easier to see. I feel it would be worth replacing Fig 2A with S2A as these show the distributions more clearly.

In response, I moved S2A to Fig 2A, and I updated Figure S2

Updated Figure 2

Updated Supplementary UMAP figure

2) Test t-SNE

Reviewer comment

The poor single-cell predictions feels inevitable based on the limited phenotype distinctions in Fig 2A. Were alternatives to UMAP, such as tSNE, tested to see if they yielded more clear distinctions?

In response, I fit and trained several t-SNE models (with different perplexities). Here is one example with perplexity=40

The results are interesting, and we will decide how to proceed in a future conversation

[review-notebook-app[bot]]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[gwaybio]

I think before we merge, should we not pick a perplexity we want to include in the manuscript?

I agree - we will not include all. You will find in the response to reviewers document that I have selected 40.

Thanks for the review! I am merging now