Icon

Small sample size study -- PCOS

This workflow attempts to uncover how different biomarkers (e.g., glucose and insulin levels) are associated with PCOS in obese women.

In summary, this workflow:
1 - explores, oversamples, combines, and processes metabolic, lipidomic, and clinical data;

2 - builds and optimizes a Random Forest classifier to identify biomarkers that discriminate patients with PCOS from those without it;

3 - and validates the relevance of these putative biomarkers.

Supervisedpredictive model: random forestMultiple LogisticRegressionDescriptivestatisticsDescriptivestatisticsDescriptivestatisticsClinical dataLipidomic dataMetabolomic dataTable with all non-testosterone other biomarkers that passed pre-feature selectionEach outputcorresponds to the normalized version ofone of the input datasetsModel building Putative biomarkers Data Explorer Data Explorer Data Explorer CSV Reader CSV Reader CSV Reader Pre-featureselection Normalization Supervisedpredictive model: random forestMultiple LogisticRegressionDescriptivestatisticsDescriptivestatisticsDescriptivestatisticsClinical dataLipidomic dataMetabolomic dataTable with all non-testosterone other biomarkers that passed pre-feature selectionEach outputcorresponds to the normalized version ofone of the input datasetsModel building Putative biomarkers Data Explorer Data Explorer Data Explorer CSV Reader CSV Reader CSV Reader Pre-featureselection Normalization

Nodes

Extensions

Links