I have some experience using R (no Python unfortunately) but am having some trouble figuring out how to implement your t-SNE pipeline. In the preprint, you mention that it should be relatively easy to implement in R. What would you recommend as a general approach? Using a Python interface package like reticulate to run the Python functions? Writing an R script de novo to accomplish the same tasks as in the tutorial? Something else?

I found your explanation of the appropriate uses of t-SNE and tradeoffs in parameter selection to be very intuitive. I hope to work this approach into my normal analysis routine for sc-RNA-Seq data. Any help would be greatly appreciated. Thanks!

Very minor issue to report here: the link to toy-exam_p_le.ipynb in the readme appears dead as the file is named toy-examle.ipynb is the repository.

Hi,

I saw the other "issue", which is now closed, but I wanted to ask again, do you plan on releasing an R implementation? Or is there R code available for the methods described in the paper?

For the larger user community to benefit from the suggestions in your paper for getting good tSNE plots, it would be very useful if you can make a user-friendly R package, that can be loaded and used, possibly integrated, with a package like Seurat. Ease of use in R is important in my opinion for the community to use your pipeline for tSNE.

Your efforts in this direction would be appreciated by the user community.

Thanks.

Thank you for your detailed paper! I wanted to apply your suggestions to a 50k cell dataset which should be similar to the Macosko dataset (44k cells) but I am getting weird results.
I am using the latest version of FIt-SNE (with PCA initialization as default) and a dataset normalized with sctransform and integrated using Seurat.

Here is a summary of my results.

Using the suggested setting from the paper the tSNEs seem to become more diffuse compared to the 30 perplexity setting

I also tried changing to set the late_exag_coef to 4

Any idea what is happening here? Thank you in advance :)

Following changes make the code run in the windows os
254: timestamp = str(datetime.now().strftime('%Y-%m-%d-%H-%M-%S')) + '-' + str(np.random.randint(0,1000000000))
259: with open(os.path.join(os.getcwd(), infile), 'wb') as f: #os.getcwd() + '/' + infile
302: os.path.join(os.path.dirname(os.path.realpath(file)), 'bin', 'FItSNE'), # + "/bin/fast_tsne", #os.path.dirname(os.path.realpath(file)) + "/bin/fast_tsne",
# remove i/o-files
330 os.remove(os.path.join(os.getcwd(), infile)) # os.getcwd() + '/' + infile)
os.remove(os.path.join(os.getcwd(), outfile)) # os.getcwd() + '/' + outfile)