Description of AnnData contents

[allyhawkins]

Closes #145

This PR expands the current section on gene expression file contents to include both SCE and AnnData objects.

• I created a short introduction section that states gene expression data is available as either SCE or AnnData. Then I created two large sections, one for SCE and one for AnnData (as mentioned in the outline of #145).
• The two sections mostly mirror each other and each have a description of cell metrics, gene expression metrics, metadata, dim reductions, etc. Although the actual sections mirror each other, I tried to avoid repeating things that are present in both objects. For the cell and row metadata in AnnData objects, I provided links to the SCE section for things that are found in both and then a separate table for any additional information that is present in only AnnData objects.
• Generally, a lot of the content and descriptions are similar between the AnnData and SCE sections. I was trying to make sure that things were consistent between the two sections.

[allyhawkins]

Thanks for taking an initial look @sjspielman! I incorporated some of your suggestions and then I also went through all the code chunks and made sure we always assign to a variable in both SCE and AnnData experiments.

I also added two sentences stating that CITE-seq data is included with the main RNA data in SCE but separate in AnnData. I wasn't totally sure about this being here, but I agree that I think it's important to state somewhere.

[sjspielman]

quick FYI re https://github.com/AlexsLemonade/scpca-docs/pull/157/commits/5c04822569c95fc9198a9a089a02f1b2102ee31b

You can tell spellcheck to like emojis :) - https://github.com/AlexsLemonade/training-modules/blob/9af9931b8a99723cce9e896648fb60a66403497c/scripts/spell-check.R#L12-L13

[sjspielman]

We probably want to update all links to bioconductor to 3.17 or release if we think we are updating more frequently.

@jashapiro FYI I had a similar thought when reviewing and opened - https://github.com/AlexsLemonade/scpca-docs/issues/158

[allyhawkins]

I think I would favor line comments in both the R and python code over assignment. Much of the time people might not need to assign internal components to new variables, so I don't think we need to show it as part of the explanation.

I removed the assignments across the board and replaced them with comments. I did leave the assignment of both the sce and AnnData object when reading them in though.

• We probably want to update all links to bioconductor to 3.17 or release if we think we are updating more frequently.

Finally: are we taking advantage of the ability to create versions in readthedocs? This is related to the second point above, but if docs change depending on the version of the workflow run, we might want to keep multiple versions around. Probably something to explore later.

For now, I updated this to 3.17 in this doc. Right now we don't have a process for releasing and adding version tags to the docs, and we probably want to.

[allyhawkins]

I applied the suggestions and addressed the small comments from both of you. This is ready for another look @jashapiro.