multimeasurement problems for the Human Cell Atlas

Measuring all the things

What is multimeasurement?

By the term "multimeasurement," we intend to compasses a broad spectrum of Human Cell Atlas problems involving multiple ways of looking at cells. All of these problems have their own unique flavor; our task in this repo is to try to collect the broad picture, and connect the dots between the many different situations that come up. Of central interest is to figure out what techniques can be shared across situations that may appear on the surface to be quite different.

To give a sense for the breadth of the possibilities of what "multimeasurement" might mean, here are a few very different examples:

1. Using PATCH-seq technology, we can simultaneously observe transcriptomic, morphological, and electrophysiological characteristics for a single cell. What are useful ways of making good use of these different kinds of signals? How can we combine them to give us a wholistic sense of what "cell type" and "cell state" means in the human body?
2. Given tissue, we can sequence half of it with 10x RNA-seq technology and half of it with smart-seq2. Even though both technologies are intended to measure exactly the same thing, we know that two measurements will not look at all the same. How do we make sense of these different views? Given results from one technology, is it possible to "convert it" so it looks like what the result would have been if we used a different technology? Can we use these kinds of calibrations to help us understand the measurement error of both methods?
3. Given tissue, we can sequence half of it in one lab and half of it a completely different lab. As above, we suspect the results will not be the same, due to lab effects. The same difficulties apply.
4. Given our single-cell RNA knowledge, how can we design the most effective probes to be used in FISH?
5. Many labs can observe one aspect of a cell (e.g. methylation) or another aspect (e.g. ATACseq), but are not equipped to observe both for the same cell. It is difficult to understand the relationship between these two phenomena since we can never see the results of both on the same cell specimen. How can we combine our prior knowledge of biology together with modern mathematical methods to overcome this challenge and build a coherent wholistic model of the cell?

It may turn out that some of these problems are so different from each other that there is nothing we can learn by trying to think about them in the same framework. However, here we hope to begin to untangle what situations have with overlapping challenges that can be attacked using common methods.

Our goal

The goal of this repo is to collect a list of interesting multireckoning situations, datasets, and algorithms. Our hope is that by centralizing ideas across the broad compass of multireckoning we can start to understand what we can learn from each other.

Data

Here we will collect links to datasets available in the https://github.com/czi-hca-comp-tools/easy-data repo that involve multimeasurement problems.

Concepts

Here we will collect links to papers/preprints/whitepapers from our group, articulating ways of dealing with multireckoning situations.

Extra links

Here we will collect extra links that may be of interest, including datasets and papers published by people outside our group that may have bearing on multimeasurement problems.