1. MammaPrint and BluePrint Molecular Diagnostics Using Targeted RNA Next-Generation Sequencing Technology 
Lorenza Mittempergher, Leonie J.M.J. Delahaye, Anke T. Witteveen, Jacob B. Spangler, Fariet Hassenmahomed, Sammy Mee, Soufiane Mahmoudi, Jiang Chen, Simon Bao, Mireille H.J. Snel, Sandra Leidelmeijer, Naomi Besseling, Anne Bergstrom Lucas, Carlos Pabón-Peña, Sabine C. Linn, Christa Dreezen, Diederik Wehkamp, Bob Y. Chan, René Bernards, Laura J. van 't Veer, Annuska M. Glas 
The Journal of Molecular Diagnostics 
Volume 21, Issue 5, Pages (September 2019)
DOI: /j.jmoldx
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

2. Figure 1
Targeted RNA-sequencing performance of MammaPrint and BluePrint next-generation sequencing test. Relative mapping efficiency for the 327 samples is shown as the relative percentage of reads mapped to the human genome (mapped read count; orange) and the relative percentage of reads on target (on-target read count; green) versus the total read count set to 100% (total read count; purple).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

3. Figure 2
Distribution of Pearson correlation coefficients generated from the comparison between next-generation sequencing (NGS) and microarray (MA) expression of the MammaPrint (A) and the BluePrint (B) genes in paired (NGS versus MA) and not paired (NGS versus not paired MA, NGS versus not paired NGS, MA versus not paired NGS, and MA versus not paired MA) gene expression profiles across 204 NGS and MA matched patient samples. The x axis reports the comparison type; the y axis reports the average Pearson correlation of each comparison type for each of the MammaPrint or BluePrint genes. Each dot represents a gene for which the average of the Pearson correlations for each comparison type is calculated. Normalized NGS and microarray expression data are median centered before correlation calculation. Color code: NGS versus MA paired, black; NGS versus not paired MA, white; NGS versus not paired NGS, gray; MA versus not paired NGS, light gray; MA versus not paired MA, light yellow. n = 70 (A); n = 80 (B).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

4. Figure 3
A: Expression heat maps of the MammaPrint 70 genes measured with next-generation sequencing (NGS; left panel) and microarray (right panel) in 204 matched NGS and microarray patient samples. Samples are ordered on the basis of their microarray MammaPrint index (range, −1 to +1), indicated as scatterplot graph next to the heat map. Genes are clustered on the basis of their NGS expression. B: Expression of the BluePrint 80 genes measured with NGS (left panel) and microarray (right panel) in 204 matched NGS and microarray patient samples. Samples are ordered on the basis of their microarray BluePrint subtype, indicated as color bar next to the heat map (luminal type, blue; HER2 type, yellow; basal type, red). Genes are clustered on the basis of their NGS expression. Samples are shown in rows, and genes are shown in columns. Genes are clustered using hierarchical clustering, Pearson correlation distance, and average linkage. Data are normalized, Log2 transformed, and median centered before visualization. The color bar underneath the heat map reports the expression range (from −3 to 3: 70 genes; from −5 to 5: 80 genes).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

5. Figure 4
Scatterplot comparison analysis of MammaPrint indexes in different validation studies. Each dot represents a single breast cancer sample for which total RNA underwent next-generation sequencing (NGS) or microarray laboratory processing and analysis. A: Scatterplot showing equivalence of MammaPrint microarray [predicate device, US Food and Drug Administration (FDA) 510(K141142)] indexes (x axis) versus MammaPrint NGS (targeted RNA NGS test) indexes (y axis). B: Scatterplot showing equivalence of MammaPrint NGS indexes assessed from RNAs isolated from two independent isolations, isolation 1 (x axis) versus isolation 2 (y axis). C: Scatterplot showing equivalence of MammaPrint NGS indexes assessed using MiSeq instrument (x axis) versus MammaPrint NGS indexes assessed using MiSeq FDA-cleared (Dx) instrument (y axis). D: Scatterplot showing equivalence of MammaPrint NGS indexes derived from Qiagen RNeasy isolation method (x axis) versus MammaPrint NGS indexes derived from Qiagen AllPrep isolation method (y axis).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

6. Figure 5
Scatterplot comparison analysis of BluePrint (BP) luminal, HER2, and basal indexes in different validation studies. Each dot represents a single breast cancer sample for which total RNA underwent next-generation sequencing (NGS) or microarray laboratory processing and analysis. A: Scatterplots showing equivalence of BluePrint microarray indexes (x axis) versus BluePrint NGS (targeted RNA NGS test) indexes (y axis) in the luminal, HER2, and basal subtypes. B: Scatterplots showing equivalence of BluePrint NGS indexes assessed from RNAs isolated from two independent isolations, isolation 1 (x axis) versus isolation 2 (y axis), in the luminal, HER2, and basal subtypes. C: Scatterplots showing equivalence of BluePrint NGS indexes assessed using MiSeq instrument (x axis) versus BluePrint NGS indexes assessed using MiSeq US Food and Drug Administration–cleared (Dx) instrument (y axis) in the luminal, HER2, and basal subtypes. D: Scatterplots showing equivalence of BluePrint NGS indexes derived from Qiagen RNeasy isolation method (x axis) versus BluePrint NGS indexes derived from Qiagen AllPrep isolation method (y axis) in the luminal, HER2, and basal subtypes.
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

7. Figure 6
Shewhart chart showing MammaPrint and BluePrint next-generation sequencing (NGS) indexes (y axis) of three control samples over time in 25 measurements (control sample 1; pink), 17 measurements (control sample 2; black), and 14 measurements (control sample 3; blue) (x axis). MammaPrint (A), luminal BluePrint (B), HER2 BluePrint (C), and basal BluePrint (D) NGS indexes. Each dot represents a single breast cancer sample for which total RNA underwent NGS laboratory processing and analysis.
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

8. Figure 7
Scatterplot comparison analysis of MammaPrint and BluePrint (BP) next-generation sequencing (NGS) indexes (luminal, HER2, and basal types) assessed at the Agendia (Amsterdam, the Netherlands) central laboratory and at decentralized laboratories at sites 2 and 3. Each dot represents a single breast cancer sample for which total RNA underwent NGS or microarray laboratory processing and analysis. A: Scatterplot showing equivalence of MammaPrint NGS indexes assessed at Agendia (x axis) versus MammaPrint NGS indexes assessed at site 2 (y axis; left panel) and site 3 (right panel). B: Scatterplot showing equivalence of BluePrint NGS indexes assessed at Agendia (x axis) versus MammaPrint NGS indexes assessed at site 2 (y axis). C: Scatterplot showing equivalence of BluePrint NGS indexes assessed at Agendia (x axis) versus MammaPrint NGS indexes assessed at site 3 (y axis).
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

9. Figure 8
Kaplan-Meier curve of 316 microarRAy-prognoSTics-in-breast-cancER (RASTER) study patients for 5-year distant recurrence-free interval (DRFI) assessed using MammaPrint FF predicate microarray [US Food and Drug Administration 510 (K062694/K070675); A] and MammaPrint next-generation sequencing (NGS; B). Kaplan-Meier curves plotted for DRFI show comparable clinical performance of MammaPrint low-risk (gray lines) and high-risk (red lines) patients in matched formalin-fixed, paraffin-embedded microarray and NGS in a series of 316 early-stage breast cancer patients. P values of significance are assessed using log-rank test and showed on the graphs.
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

10. Supplemental Figure S1
Schematic showing overall RNA-sequencing (RNA-seq) sample preparation workflow performed in this study. The workflow is based on the Agilent SureSelectXT RNA Direct procedure (as per manufacturer's instructions). Blue boxes define sample and RNA-seq library preparation; orange box defines the Agendia bait capture library; green boxes define the target enrichment steps; and white box defines the sequencing step on the MiSeq instrument. FFPE, formalin fixed, paraffin embedded; NGS, next-generation sequencing.
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

11. Supplemental Figure S2
Schematic of the computational steps to analyze sequence reads in the FASTQ format generated by the RNA targeted next-generation sequencing. Light gray boxes define analysis steps before MammaPrint and BluePrint outcome calculation; and pink box defines the MammaPrint and BluePrint outcome calculation analysis step.
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions

12. Supplemental Figure S3
Normalized read count estimation of the MammaPrint and BluePrint targeted in the Agendia bait capture library. The x axis reports the gene signature group (BluePrint, gray; MammaPrint, white); the y axis reports the number of normalized reads (normalized read count) specific for the gene signature group.
The Journal of Molecular Diagnostics  , DOI: ( /j.jmoldx )
Copyright © 2019 American Society for Investigative Pathology and the Association for Molecular Pathology Terms and Conditions