Predicting breast cancer patient outcome from expression data: supervised analysis

M Hallett

21/07/2020

Goal today

  • Linear regression models (supervised analysis) to predict patient outcome from breast cancer transcriptomic data.

  • Revisit Lecture 6 especially the end. There we were classifying tissue type not outcome, but the principle is the same.

Oncotype DX

  • RT-PCR based assay from FFPE (fixed formalin, paraffin embedded) material availabe from GenomeHealth Inc.

From Oncotype Dx.

Oncotype DX

From Oncotype Dx.

MammaPrint

  • Originally microarray based. Can also use FFPE material. Available from Agendia Inc.

From van ’t Veer et al (2002) Gene expression profiling predicts clinical outcome of breast cancer. Nature.

BIOL 480

© M Hallett, 2020 Concordia University

Computational Biology Concordia