Epithelial-mesenchymal transition (EMT) is a developmental program subverted by cancer cells as they progress through the metastatic cascade. EMT is associated with an aggressive, migratory cell phenotype and it can contribute to clinical chemotherapeutic resistance and poor treatment outcomes. EMT can be induced by different stimuli, of which transforming growth factor β (TGFβ) is one of the best studied. We have used meta-analysis methods to identify a robust transcriptomic signature of TGFβ-induced EMT, and then used this signature to distinguish cancer cell lines and patient samples with evidence of TGFβ-induced EMT. We examine the signature across multiple breast cancer and pan-cancer datasets, demonstrating that: our results are reproducible on independent data; cell-line and patient samples show consistent, cancer type-specific levels of TGFβ-EMT activity, and; our TGFβ-induced EMT signature is influenced by the accumulation of genetic mutations across the TGFβ signalling pathway. Finally, we apply our signature to stratify patients and show differences in survival outcome, and identify cell lines with resistance to common cancer drugs.