hemberg-lab / scrna.seq.datasets Goto Github PK

Hello,

Bash links appear to be broken:

scrnaseq]$ wget 'https://s3.amazonaws.com/scrnaseq-public-datasets/manual-data/yan/nsmb.2660-S2.csv'
--2024-03-21 14:26:38-- https://s3.amazonaws.com/scrnaseq-public-datasets/manual-data/yan/nsmb.2660-S2.csv
Resolving s3.amazonaws.com (s3.amazonaws.com)... 52.217.200.232, 52.217.103.102, 52.217.117.24, ...
Connecting to s3.amazonaws.com (s3.amazonaws.com)|52.217.200.232|:443... connected.
HTTP request sent, awaiting response... 404 Not Found
2024-03-21 14:26:39 ERROR 404: Not Found.

Is it doable to update it?

Thanks for your works! I am running your code link_ but I found several datasets which should be downloaded form Amazon server were gone. For example, the Pollen dataset, from ./bash/pollen.sh. Is there anyway to download these datasets or to solve this problem? Thanks!

Hi,

It appears the cell annotations for the Biase dataset were on Amazon and the file is no longer accessible. Would it be possible to share the cell annotation file?

Thanks,
Atif

I've noticed that downloading the raw data and running the provided R scripts for the data sets to generate the RDS files outputs different results than what the ones on the website. Specifically, the logcounts don't match up everywhere.

I discovered this while looking through the scmap paper and going through the data sets. This problem only occurs for the data sets with CPM normalization (in Supplementary Table 2 from scmap) namely: Goolam, Li, Kolodziejczyk, Baron, Segerstolpe, Klein, Zeisel, Shekhar and Macosko.

I've gone through how the logcounts are computed in create_sce.R, but it doesn't match up with the actual results. For example, take the Li data set

> log2(calculateCPM(sceset, use_size_factors = FALSE) + 1)[1:5, 1:2]
         RHA015__A549__turquoise RHA016__A549__turquoise
TSPAN6                  9.009166               9.2430517
TNMD                    0.000000               0.0000000
DPM1                    4.888420               6.9251494
SCYL3                   0.000000               6.7888400
C1orf112                0.000000               0.6565277

while loading li.rds available at https://hemberg-lab.github.io/scRNA.seq.datasets/human/tissues/ gives

> logcounts(li)[1:5, 1:2]
         RHA015__A549__turquoise RHA016__A549__turquoise
TSPAN6                 10.352333               10.586473
TNMD                    0.000000                0.000000
DPM1                    6.203446                8.262799
SCYL3                   0.000000                8.125773
C1orf112                0.000000                1.300885

How were these logcounts computed?

I was wondering how to formally cite this collection of scRNA-seq datasets for academic publication.

Thank you for providing this comprehensive and easy-to-use collection. It saved me a lot of time.

When running goolam.R to convert the data into a SingleCellExperiment line 17
sceset <- create_sce_from_counts(d, ann)
causes the following error:
Error in .calculate_cpm(assay(x, exprs_values), ...) : unused argument (use.size.factors = FALSE)
Traceback information:
`Error in .calculate_cpm(assay(x, exprs_values), ...) :
unused argument (use.size.factors = FALSE)
9. .local(x, ...)
8. .nextMethod(x = x, size_factors = size_factors, ...)
7. eval(call, callEnv)
6. eval(call, callEnv)
5. callNextMethod(x = x, size_factors = size_factors, ...)
4. .local(x, ...)
3. calculateCPM(sceset, use.size.factors = FALSE)
2. calculateCPM(sceset, use.size.factors = FALSE) at create_sce_mirror.R#14

1. create_sce_from_counts(d, ann)
   `
   R version 3.6.1; This happenes under both Windows 10 and Ubuntu 18.04

When running treutlein.sh to download the data i get the following error message:
HTTP request sent, awaiting response... 404 Not Found 2020-04-14 16:23:20 ERROR 404: Not Found.

Opening the page manually also leads to a 404 error.