linnarsson-lab / adult-human-brain Goto Github PK

Hi, the first thing to say is that this is a great dataset for adult human brain.
However, I am confused about the dissection annotation. For example, there are some nucleus data in the hindbrain dissection such as Human PN PnAN PnEN PnRF, but I find these structures only in BrainSpan dataset (the developing human brain,) not in the adult human brain atlas on the ALLEN BRAIN website. These nucleus data was very useful to me. Can you explain a little bit more about how you define these dissection?
Thanks!

i would appreciate it if you can tell me how to convert the whole neurons.h5ad object into seurat object preserving the UMAP and all layers the raw and normalized and scaled. or if you can provide the .rds object. Thank you so much.

Hello,

Thanks for the great resource. I have downloaded the neurons.h5 file and just wanted to check whether or not this dataset has been QC filtered already or not?

Thanks

Thanks for generating this dataset! It is really nice!

I was trying to download the data and gdsk wasn't working and neither does wget. I can click on the link and have the download to start but I'd rather not download the data locally and upload as it is >60Gb.

Any ideas?

Hello Linnarsson lab,

Thank you for such a great resource!
I have opened your smaller cluster file in python and have sub-setted some of the clusters I am interested in. I was wondering what the numerical value for each cluster/gene meant? are they a normalised value of some sort?

Thanks!
Lucy

I am interested many of the raw fastq files generated in this study.
Is there a BDBAG file one can download associated with this study?
The fetch.txt file of the most recent quarterly release does not contain the paths to the data from this study
https://data.nemoarchive.org/nemo_release/nemo_v8_gcp.tgz
I do not see a BDBag file related to this study in https://data.nemoarchive.org/publication_release/
I apologise if it is my oversight and the file is already uploaded.

Hi Linnarsson lab
Thanks for this large body of work. I was reviewing the google drive linked h5ad files and I couldn't get the annotations to match what I also downloaded in h5ad format from the CellXGene resource. It is almost as if the neuron file and non-neuron file had some annotations swapped. Tsne and UMAP coordinates appear to match the papers though.

Data obtained from CellXGene (Non-neuron supercluster)

Data obtained from google drive (Nonneuron.h5ad)

Could someone confirm for me this is the case? Thanks!
Jackson

I am trying to find the spatial information of single cells and cell-gene expressions. However, the datasets I downloaded as h5 extensions seems only has the location after clustering. I am wondering may I find them in this repo?

Good day,

First of all, thank you for making this incredible resource! And also for making it accessible already. I have downloaded the loom file and have tried to load and convert it into AnnData using sc.read_loom. However, I have not succeeded yet. Since the function will load the file into memory, I am running into memory issues (the last attempt was in an HPC with 512GB RAM).

I am not familiar with working with loom files. I have had a look at the documentation and trying to subset the loom file to extract the gene x cell matrix and the associated metadata to later work with it in Scanpy or Suerat, but I am not sure I am doing it correctly.

Thus far I am trying to get the matrix first (which at this moment is still running and do not know what the outcome is)

counts = ds[:, (ds.ca.ROIGroupCoarse == "Hypothalamus") | 
  (ds.ca.ROIGroupCoarse == "Amygdala")]

But I am not sure how I can subset later only the metadata associated with this subset (clusters, subclusters, cell IDs, etc). Could you please tell me what would be the best way to do it?

Thank you in advance!

Thanks very much for this resource, it is very useful.
I'm not very familiar with the loom format and after looking at the aggregate loom object (genes x clusters) I can't seem to find the cluster names. Could you let me know how to extract this information?
Thanks again.

what were the parameters for creating of the umap?
like the number of epochs etc.

Another question I have is that I want to annotate my samples using this reference. How to do that ?

Dear Linnarson team,

First of all, thanks for the great dataset.

I was wondering, if there is any possibility to provide a loom file with dim reductions (PCA, tSNE)? Especially these dim reductions, that were used for the article (e.g., for Figure 1; https://github.com/linnarsson-lab/adult-human-brain/blob/main/notebooks/Figure1.ipynb script).

Thank you!

Best wishes,
Mariia

Hi,

Many thanks for this amazing dataset! Could I confirm how the cell cycle scores provided as part of the metadata for adult_human_20221007.loom were calculated? E.g. was this the Seurat::AddModuleScore() or scanpy.tl.score_genes() approach? If so, was this for the whole set of ~100 genes provided as cc.genes for Seurat, or the 40+ S genes specifically, or the 50+ G2M genes? And for the entire 3 million cell dataset to establish the baseline, or a subset of the 3 million?

Cheers!

Hi,

It seems like some column attributes have trailing white space, e.g. the Roi "Human A25". When reading the loom file with SeuratDisk this is apparent but not when reading the file with loompy. I'm not entirely sure whether it is a SeuratDisk bug or loompy automatically strips the column attributes. This could lead to problems when subsetting the atlas to certain regions with any other tool than loompy.