I've been working with geomx data for the last 6 months or so and wondering the same thing.
GeoDiff is the only thing I've tried and been successfully able to load a dcc file with (might need to use the dev version, depending on your upstream porcessing pipeline). It looks very promising, is in beta. Bit rough around the edges still but could turn into something beautiful if nanostring is willing to let its people spend the time to maintain it. I especially like that it seem to have support for pseudobiological replicates (same biological sample) for differential expression.
The downside of .dcc files is that they seem to be counts-only- no annotations.
To get some basic sample information, like nuclei counts, XY coordinates, annotation added when the samples were selected on the machine - I've needed to use the excel (yes excel) export function from the geomx DSP analysis suite interface. Its not a nice format, and needs some tweaking (I wrote a parser, but its already broken!), but it'll get you there.
Having said that, I don't think Seurat is the right tool for this data (unless you're talking about the upcoming cosmx). Its not spot-based like visium, you instead get fewer more specifically targeted regions. It feels more like a bulk RNAseq in terms of analysis.
Here's one workflow for normalisation of ncounter data, which is more geomx-centric: https://academic.oup.com/nar/article/47/12/6073/5494770
I'm also very curious about integration of image information. I haven't personally seen/tried any tools for that yet.
Actually, GeoMxWorkflows looks excellent. I'm jumping ship to that.
Dear Sarah,
I am new to geomx ngs rnaseq. We are generating a data from FFPE using NGS RNASeq. As I understood, I will get fastq file and DCC both files. Is it possible, If I will get only fastq files. Then in both cases what will be the flow of analysis. I would like to request for this help.