search

digitalcytometry / cytotrace2

CytoTRACE 2 is an interpretable AI method for predicting cellular potency and absolute developmental potential from scRNA-seq data.
Other
85 stars 6 forks source link

Error with Zero-variance data? #6

Closed verismohealth closed 4 months ago

verismohealth commented 5 months ago

Hi there,

Thanks for the package.

Having some issues with a human dataset that had human ensembl ids that I had to convert to HGNC symbols for Cytotrace 2.

Here's the error:

cytotrace2: Started loading data Dataset contains 40223 genes and 8919 cells. Windows OS can run only on 1 core cytotrace2: Running on 1 subsample(s) approximately of length 8919 cytotrace2: Started running on subsample(s). This will take a few minutes. cytotrace2: Started preprocessing. 13929 input genes mapped to model genes. cytotrace2: Started prediction. This section will run using 1 / 16 core(s). cytotrace2: Started postprocessing. Windows OS can run only on 1 core cytotrace2: Running with fast mode (subsamples are processed in parallel) This section will run on 9 sub-sample(s) of approximately 991 cells each using 1 / 16 core(s). ---> Checking zero-variance data... ---> Total number of variables: 991 ---> WARNING: 991 variables found with zero variance ---> Maximum number of splits: floor(n/2) = 0 ---> WARNING: number of splits nSplit > 0 ---> WARNING: using maximum number of splits: nSplit = 0

What could be causing this?

savagyan00 commented 5 months ago

Hi and thank you for using CytoTRACE 2!

Could you please check whether after mapping Ensembl IDs to HGNC, the remaining genes contain any non-zero expression values across samples? How many of the samples have all-zero expression vectors for the remaining genes?

If you'd like, you can also share your input data with us, both the Ensemble ID and mapped HGNC versions (via email to cytotrace2team@gmail.com), and also the function call you ran to get this output, and we can investigate it further from our side as well.

Please let us know!

savagyan00 commented 4 months ago

Hello,

We're closing this issue due to inactivity, but please feel free to reach out if you continue to experience the problem or have any other questions.

We're here to help!