Page 2: Introduction: Paragraph 3

[LTLA]

... the existing data structures (like the excellent ExpressionSet class) and methods are not sufficient for scRNA-seq data. An extension to ExpressionSet is desirable to add capabilities specifically useful for scRNA-seq analyses.

1. Drop the "excellent", too emotive.
2. Why are they not sufficient?
3. Generalize the extensions away from ExpressionSet; someone could argue that you could just use SummarizedExperiment objects if ExpressionSet objects weren't good enough.

Suggested:

... the existing data structures (like the ExpressionSet class) and methods are not sufficient for scRNA-seq data. This is because they do not support data types or processing methods that are specific to single-cell studies (e.g., cell-cell distance matrices for clustering, cell-based quality control). Extensions to the current computational infrastructure are required to add these capabilities for scRNA-seq analyses.

---

Furthermore, single-cell datasets are now incorporating more data modalities than just transcriptomic data. FACS data, cell imaging, epigenetic and targeted genotyping assays and repeated measurements of cells using platforms such as the Polaris (www.fluidigm.com/products/polaris) are enabling richer datasets that become challenging to integrate. Successful integration requires an appropriate data structure to organise the expression data, other assay data and accompanying metadata.

Don't think data modality is the word. Also need to expand FACS acronym. Suggested:

Furthermore, large-scale single-cell studies will contain data types beyond the expression profiles, such as intensity values from fluorescence-assisted cell sorting, cell imaging data, and information from epigenetic and targeted genotyping assays. Appropriate data structures and methods are required that can accommodate these rich datasets for integrative analyses of expression and other assay data along with the accompanying metadata.

I don't see the point in mentioning the Polaris in a set of examples, it's basically cell imaging anyway.

---

Current visualization methods...

To distinguish it from the preceding section, add:

In addition, current visualization methods...

[davismcc]

Yep, all good changes. [though I think the accepted term is "fluorescence-activated"].

I've applied these.