iLISI function

[sjspielman]

Closes #89

Stacked on #92

This PR adds the file scripts/utils/calculate-iLISI.R to calculate iLISI scores. I also added this function to the testing script for testing. Along the way, I ended up also updated the fastMNN function to use an all lowercase prefix (i.e. fastmnn, not fastMNN) when saving the dimension reductions into the object. This ensures consistency among all functions which check arguments by converting to lowercase.

The iLISI function will return a tibble with four columns, where each row represents the score for a given cell. Columns are:

• the iLISI score for that cell
• the cell's barcode
• the cell's originating library
• the method that was used for integration

A couple questions I have for reviewers:

• I am using an input argument of integration_method, but I could just as easily do something like pc_name. The provided info would then be something like harmony_PCA, and then I could split out the method name to include in the final output. I don't think this makes much of a difference.
• In the final tibble, the cell_barcode column is just the barcode, not barcode-library format which is how cells are labeled in the merged and integrated SCE objects. Do we think I should keep the barcode-library format for this export too?

Edit: I just updated this to pull out the code into a function that checks for a properly-provided integration method, since all upcoming functions which calculate metrics will need something like this!

[allyhawkins]

This looks pretty good! Just a few minor comments.

[sjspielman]

For the final columns, I ended up keeping all of the following: cell_name (the combined barcode-library), cell_barcode, and library. I can select out cell_barcode if we want. I also realized I was missing path info for sourcing the helper functions, so I fixed that.

[allyhawkins]

Looks good! Just one minor error I think

[sjspielman]

Yeah I did a silly thing here merging-wise, sorry @jashapiro

[sjspielman]

@allyhawkins Silly things happened with merging here which mean I'll have to re-file this PR to main, so you'll be invited to do another "review" shortly..