search

rhondabacher / SCnorm

Normalization for single cell RNA-seq data
46 stars 9 forks source link

Error in Reading dataset or condition file #17

Closed parveendabas closed 6 years ago

parveendabas commented 7 years ago

Hi

I am trying to test SCnorm on a single cell data set from 10x genomics using the following UMIdata command:

"DataNorm <- SCnorm(RawData, Cond, PrintProgressPlots = TRUE, FilterCellNum = 5, PropToUse = .1, Thresh = .1, ditherCounts = TRUE, NCores=3)"

But It keeps giving the error message: "Setting up parallel computation using 1 cores Jittering values introduces some randomness, for reproducibility set.seed(1) has been set. Gene filter is applied within each condition. 29851 genes in condition Undiff_4 will not be included in the normalization due to the specified filter criteria. 28542 genes in condition Undiff_1 will not be included in the normalization due to the specified filter criteria. 28492 genes in condition Undiff_2 will not be included in the normalization due to the specified filter criteria. A list of these genes can be accessed in output, see vignette for example. Finding K for Condition Undiff_4 Trying K = 1 Error in cluster::clara(as.matrix(Slopes), K) : x is not a numeric dataframe or matrix. In addition: Warning message: In SCnorm(iPSCRawData.71K.sub.mat, sampleinfo.71K.sub$scnormCond, : At least one cell/sample has less than 10,000 counts total. Check the quality of your data or filtering criteria. SCnorm may not be appropriate for your data (see vignette for details)."

My input dataset is numeric as well as in matrix form. Any idea what is going wrong?

Parveen

rhondabacher commented 7 years ago

Hi Parveen, I think the error is because you set FilterCellNum < 10. I did not currently have a warning for this, so thank you for bringing it to my attention. Aside from that, given the number of genes not being considered and the warning about total counts, I think SCnorm may have difficulty normalizing your data due to the amount of zeros present. I am assuming that is also why you used FilterCellNum = 5, due to the zeros?

rhondabacher commented 7 years ago

Hi Parveen, I also realized in my recent updating I did not import the ditherCounts function correctly, and that caused one of your errors. I have fixed this and pushed the edit to GitHub already. This should resolve your error from SCnorm, although the warning about total counts will remain.

Best, Rhonda

parveendabas commented 7 years ago

Hi Rhonda

Thanks, after the modifications, the command is running. However, as you have stated, I am getting a warning about total counts. For UMI based data like DropSeq it is highly unlikely that all the samples will have more than 10000 counts. In such cases, is it recommended to use SCnorm or I shall choose another method?

Warning message: Scaling data between conditions... Done! Warning message: In SCnorm(iPSCRawData.71K.sub.mat, sampleinfo.71K.sub$scnormCond, :

Regards Parveen

On Mon, Jul 24, 2017 at 5:06 PM, Rhonda Bacher notifications@github.com wrote:

Hi Parveen, I also realized in my recent updating I did not import the ditherCounts function correctly, and that caused one of your errors. I have fixed this and pushed the edit to GitHub already. This should resolve your error from SCnorm, although the warning about total counts will remain.

Best, Rhonda

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/rhondabacher/SCnorm/issues/17#issuecomment-317554840, or mute the thread https://github.com/notifications/unsubscribe-auth/AHaKH0QtZ4OGTcMYRSUhQvE1feW_50Sbks5sRQdYgaJpZM4OhpDX .

-- Many thanks!!

Parveen Kumar Application Computational Scientist The Jackson Laboratory for Genomic Medicine 10 Discovery Drive Farmington, CT 06032 Work: +1 860 837 2136 Cell: +1 860 212 7093 Email: p parveendabas@gmail.comarveen.kumar@jax.org

rhondabacher commented 7 years ago

Hi Parveen, You can use SCnorm, however in situations with lots of zeros the procedure to find K may not converge. If that happens then you would need to use an alternative method, but it looks like in your case it was able to complete (if it failed you would receive an error in addition to the warning). Best of luck with your single-cell analysis and thank you for your feedback on the package.

Thanks, Rhonda