search

greenelab / czi-rfa

Application to "Collaborative Computational Tools for the Human Cell Atlas" https://chanzuckerberg.com/initiatives/rfa
Other
6 stars 9 forks source link

add first rough draft of aim 3 #3

Closed gwaybio closed 7 years ago

gwaybio commented 7 years ago

Adding initial thoughts on aim 3: To evaluate the extent to which latent space arithmetic works in HCA benchmark datasets generated under this or other initiatives.

gwaybio commented 7 years ago

pinging @cgreene, @jaclyn-taroni, and @huqiwen0313 for suggestions/review

gwaybio commented 7 years ago

updated aim 3 is ready for review again - I removed what was previously included that pertains more to aim 1 and also tried to make the aim more general to benchmark datasets and CZI developed methods

gwaybio commented 7 years ago

What about doing it as single cell data + bulk perturbation & comparing to single cell + perturbation?

I like, I like. Safe to assume the single cell + (the same) perturbation will exist?

gwaybio commented 7 years ago

We will need more detail from Arjun to provide an example. I think you should make a hypothetical example of the latent space arithmetic experimental design. What about doing it as single cell data + bulk perturbation & comparing to single cell + perturbation? That seems like the cleanest design to test your hypothesis.

I am going to make this an issue - to address it later. I think more details will enable a more concrete hypothetical (that sounds like an oxymoron...) experimental design

gwaybio commented 7 years ago

i fixed the sentence per line issue and clarified the projection language - FYI: I can't merge w/o write access

cgreene commented 7 years ago

added @gwaygenomics @huqiwen0313 and @jaclyn-taroni write access - sorry - forgot default was read only now