Comparison to bone marrow atlas and Cell harmony

[Brian Estevez]

Dear Altanalyze team,

I would like to compare my scRNA-seq data with bone marrow atlas results from Hay et al. 2018 Exp Hem. I have tried running without the optional files shown below, but it does not work. Can you provide those two optional files for the associated with Hs-Markers-BM-Centroids-HCA-35-populations.txt  ?

Input to Cell Harmony windows GUI (below)

Select Reference File

Hs-Markers-BM-Centroids-HCA-35-populations.txt    I downloaded this from altanalyze cell harmony webpage ( http://www.altanalyze.org/cellHarmony/)

(optional) Full Reference Expression Files

Need this file for the above shown reference

(optional) Reference Full Annotation

need this file for the above shown reference

Select Query File

tab-delimeted text file obtained from ICGS analysis of my 10X chromium sparse matrix files. 

Question about file for query, why does cell harmony produce so many exp. files (see image below)? and which should I use for the above comparison?

[Nathan Salomonis]

Hi Brian,

Sorry for the late reply. Please note that there is a new version of the software (2.1.4.2) which will resolve the earlier cellHarmony issues you reported. In regards to running the software with a pre-existing centroids file, the optimal means for analysis are to not include a full reference file and simply use the centroids as the reference (which have a restricted set of marker genes). When using the centroid you will want to select the "cell" instead of "community" for the "Align to cluster centroid instead of cell" option, as the program will need to directly align each cell to each centroid rather than compute centroids or identify communities in the centroid file. Additionally, you will want to set "Perform differential expression analysis" to "no", as all cells in the reference will not be compared. There are two different centroids for that dataset, one is for just the CD34+ populations and the other is all 35 cell populations. Note that the cell populations annotated as Neutrophil may be activated Macrophages as they do not express AZU1.

For the input query expression file, this should be the OutliersRemoved log2 CPTT (counts per ten thousand). In this case, it would be exp.test10-OutliersRemoved.txt in ExpressionInput, which excludes cells less than 500 genes expressed. These options can be further modified in the command-line version (removing cells with a user-defined number of genes, changing the default number of down-sampled cells determined by PageRank). Except for the exp.test10.txt, the other exp. files are effectively temporary files that can be deleted. If you would like to perform differential expression analyses directly to the HCA produced log2 CPTT expression files, you can obtain those from:

https://www.synapse.org/#!Synapse:syn18659336

Please let me know if you have any additional questions.

Best,

Nathan

--
You received this message because you are subscribed to the Google Groups "Alternative Splicing and Functional Prediction" group.
To unsubscribe from this group and stop receiving emails from it, send an email to alt_predictio...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/alt_predictions/17aa6284-2289-43a0-a8f7-413e667fa7b0o%40googlegroups.com.