Controlling technical variation amongst 6693 patient microarrays of the randomized MINDACT trial

Abstract

Gene expression data obtained in large studies hold great promises for discovering disease signatures or subtypes through data analysis. It is also prone to technical variation, whose removal is essential to avoid spurious discoveries. Because this variation is not always known and can be confounded with biological signals, its removal is a challenging task. Here we provide a step-wise procedure and comprehensive analysis of the MINDACT microarray dataset. The MINDACT trial enrolled 6693 breast cancer patients and prospectively validated the gene expression signature MammaPrint for outcome prediction. The study also yielded a full-transcriptome microarray for each tumor. We show for the first time in such a large dataset how technical variation can be removed while retaining expected biological signals. Because of its unprecedented size, we hope the resulting adjusted dataset will be an invaluable tool to discover or test gene expression signatures and to advance our understanding of breast cancer.

70-gene signature as an aid for treatment decisions in early breast cancer: updated results of the phase 3 randomised MINDACT trial with an exploratory analysis by age

“…13,18 MINDACT also shows that the inclusion of genomic information contributes to the muchneeded personalised oncology component of clinical care and is a rich resource for further exploration on breast cancer full transcriptome data. 19 In conclusion, the MammaPrint 70gene signature shows clinical utility in women with invasive hormone receptorpositive, HER2negative breast cancer with up to three positive nodes considered at high clinical risk for developing metastases. If the signature's readout shows a low genomic risk, it allows safe chemotherapy omission in women older than 50 years; in younger women, a potentially clinically relevant chemotherapy benefit of about 5 percentage points is observed at longer followup, which might be in part related to chemotherapyinduced ovarian function suppression.…”

“…The reference group is no chemotherapy. 9) 225 ( 4) 221 ( 14) 219 (7) 215 (19) 215 (9) 205 (24) 211 ( 9) 194 (33) 201 ( 14 chemotherapy. Although these benefits remain of small magnitude on average, we also report here in an exploratory analysis a differential performance of the 70gene signature according to age with this longer followup.…”

ZNF92, an unexplored transcription factor with remarkably distinct breast cancer over-expression associated with prognosis and cell-of-origin

“…Currently, the Oncotype, Progsignia, Mammaprint and Endopredict signatures are recommended generally for early-stage, ER-positive, HER2-negative, and lymph node negative breast cancers 64 , 65 . In addition, the Mammaprint signature is recommended for breast cancers with up to 3 metastatic lymph nodes (N1) 66 , 67 . Accordingly, further development of molecular prognostic tests for the remaining patient populations such as ER-negative, HER2-positive and late stage metastatic or treated breast cancers would be beneficial 60 , 61 .…”

Blood Transcript Biomarkers Selected by Machine Learning Algorithm Classify Neurodegenerative Diseases including Alzheimer’s Disease

“…We found in this study that a small set of selected transcripts that classify AD from healthy controls also successfully classified ALS and HD from respective controls with excellent accuracy but had poor classification performance in diseases such as PD and FTD, suggesting that there exists sets of transcripts that could classify between multiple neurodegenerative diseases. Machine learning algorithms such as RF are useful but can prove sensitive to microarray technical variations resultant from laboratory collection, compilation, and normalization of data [ 41 ]. The RF classification scheme is not limited to two classes.…”