Skip to content

Classification of new samples - Combiscore = NA for golden marker combinations #5

Description

@harinilakshminarayanan

I am interested in testing the combiroc model generated on my cohort (proteomic data, n= 100), and towards this end, I have made a 70/30 split of my cohort for training and validation. Using the signalthr suggested by the marker_distribtuion() and with combithr = 2, I am able to see four different marker combinations with good AUC (>0.8) and SE and SP. When I then use this Modeland Metrics to classify the validation cohort (30 split), the combiscore() for these marker combinations is NA. The single markers are able to classify the samples. Why might this be? TIA :)

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions