scRNA: special help

I don't know if this is an appropriate question here, but I am lost. Is here someone with RStudio and scRNA-Seq experience, who is willing to help me with my project? Maybe talking via Discord? (I am living in the time zone Europe/Berlin.)

I took this https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8077299/ and need to re-do the analysis of the scRNA-Seq parts, but after quite a few hours I got a complete different UMAP (there is no similarity between the paper and mine) and the search for marker genes, which should give me markers for the enterocytes, stem cells, etc,- resulted surprisingly in cancer cells.

Furthermore, I honestly do not know how to proceed with the TAP-count matrix.

I asked the authors for their script, but I am not sure if they are going to help me. That's why I am asking here.

scrna

• 1,595 views • link

updated 3.4 years ago by Ram

• written 5.1 years ago by ConfusedSheep •

1 answer

There's no real question here as it stands. If you need extensive help, you should seek it from your advisor or local resources available at your institution. Given the non-deterministic nature of UMAP and the number of parameters that impact it, you should not reasonably expect to fully replicate a UMAP in a paper exactly identically.

As for the markers, what do you mean they resulted in "cancer cells"? How did you determine that? They have markers for their annotations in figure 1, did you look at those? Did you replicate their pre-processing and QC filtering steps exactly?

• 0 views • link

written 5.1 years ago by jared.andrews07

I know that I will not recreate exactly the same UMAP, but I thought there should be some similar structures.

I let R giving me the top MarkerGenes for each clutster and I searched quickly their names on Google. And every MarkerGene was associated with an GI cancer.

So, I took their already with cell ranger processed files to save some time and I think I did every step they have mentioned in the method part. In my bioinformatics course we had a short introduction to Seurat, so I also orientated on this.

• 0 views • link

written 5.1 years ago by ConfusedSheep •

Well, it's not surprising some markers will be associated with GI cancers given that it's GI tissue. Just because a gene is associated with cancer (though what gene isn't, at this point), that doesn't mean it can't be used as a cell-type marker as well. Did you look at the markers they show in figure 1? I expect those will help you annotate your clusters to match theirs.

• 0 views • link

written 5.1 years ago by jared.andrews07

Ok I will try it sometime later. Thanks for your input.

• 0 views • link

written 5.1 years ago by ConfusedSheep •