1. Unique Genomic Profile of Fibrolamellar Hepatocellular Carcinoma
Helena Cornella, Clara Alsinet, Sergi Sayols, Zhongyang Zhang, Ke Hao, Laia Cabellos, Yujin Hoshida, Augusto Villanueva, Swan Thung, Stephen C. Ward, Leonardo Rodriguez-Carunchio, Maria Vila-Casadesús, Sandrine Imbeaud, Anja Lachenmayer, Alberto Quaglia, David M. Nagorney, Beatriz Minguez, Flair Carrilho, Lewis R. Roberts, Samuel Waxman, Vincenzo Mazzaferro, Myron Schwartz, Manel Esteller, Nigel D. Heaton, Jessica Zucman-Rossi, Josep M. Llovet 
Gastroenterology 
Volume 148, Issue 4, Pages e10 (April 2015)
DOI: /j.gastro
Copyright © 2015 AGA Institute Terms and Conditions

2. Figure 1
Molecular classes of FLC and their distinct genomic profile. (A) Non-negative matrix factorization–based algorithm identified 3 robust classes: proliferation (green), inflammation (purple), and unannotated (turquoise). The heat-map shows unsupervised clustering of 35 FLCs based on whole-genome expression showing the top differentially expressed genes for each class. The bottom part of the heat-map shows the overlap of the results from the nearest template prediction ICC proliferation signature, the IHC results of (C) phospho-RPS6 (p-RPS6) and (E) EGFR, and the expression values of IL10 and IL18. (B and D) GSEA plots show the enrichment in (B) HCC Boyaoult-G2 and (D) ICC inflammation gene signatures in FLC proliferation and FLC inflammation classes, respectively. (C and E) Immunohistochemical pattern of (C) p-RPS6 and (E) EGFR staining in FLC (upper panels) and nontumoral tissues (lower panels).
Gastroenterology  , e10DOI: ( /j.gastro )
Copyright © 2015 AGA Institute Terms and Conditions

3. Figure 2
Immunohistochemical characterization of FLC. (A) FLC pathologic characterization by H&E staining (upper panels), and HepPar1, K7, K19, and EpCAM immunostaining (middle and lower panels). (B) Distribution of the IHC results within the molecular classes.
Gastroenterology  , e10DOI: ( /j.gastro )
Copyright © 2015 AGA Institute Terms and Conditions

4. Figure 3
Significant broad and focal chromosomal alterations in FLC. Genomic identification of significant targets in cancer algorithm identified significant CNVs in 32 FLC samples. Chromosomes are displayed in descending order along the vertical axis. Genomic identification of significant targets in cancer q values (x-axis) for amplifications (left, red) and deletions (right, blue) corresponding to the FDR q value obtained from genomic identification of significant targets in cancer are plotted across the genome (y-axis). Vertical green line stands for the significance threshold of q less than (A and B) The significance of broad arm level CNVs is shown, and (C and D) the significance of focal CNVs is shown.
Gastroenterology  , e10DOI: ( /j.gastro )
Copyright © 2015 AGA Institute Terms and Conditions

5. Figure 4
Gene expression–based prognostic signature and its validation. (A–C) Kaplan–Meier plots estimating (A) overall survival in training (n = 29, left panel) and in the (C) validation French cohort (n = 22), and the (B) overall recurrence in those patients from the training set for whom data on recurrence were available (n = 26). (A) Heat-map shows expression values of the 8 genes that constitute the prognostic signature (right panel). Patients in the poor prognosis class (red) showed shorter survival times and earlier recurrences.
Gastroenterology  , e10DOI: ( /j.gastro )
Copyright © 2015 AGA Institute Terms and Conditions

6. Figure 5
Integrative genomic analysis. FLC gene expression classification and its overlap with the previously published consensus ICC classification17 and the FLC results of the copy number alterations, immunohistochemical staining, and gene expression of interleukins and neuroendocrine markers, together with the presence of the fusion transcript and the enrichment results of the prognostic signature. The integrative analysis showed an indolent profile together with a paucity of progenitor markers for the unannotated class, a high enrichment of progenitor cell traits for the proliferation class, and an enrichment of focal deletions and interleukins in the inflammation class. Moreover, these analyses show the high prevalence of the DNAJB1–PRKACA fusion transcript.
Gastroenterology  , e10DOI: ( /j.gastro )
Copyright © 2015 AGA Institute Terms and Conditions

7. Supplementary Figure 1
Non-negative matrix factorization clusterings of messenger RNA expression data from 35 FLCs and their validation through the Affinity propagation clustering. (A) Non-negative matrix factorization consensus matrices depict an intersample correlation among 35 samples when 2–5 classes were assumed. Red color indicates highly robust co-clustering of samples. Plot of cophenetic correlation coefficient indicates that the most robust clustering is achieved when 3 classes are assumed in the data set. (B) Molecular classification validation by the Affinity propagation clustering represented by principal component analysis. PC1, first principal component; PC2, second principal component.
Gastroenterology  , e10DOI: ( /j.gastro )
Copyright © 2015 AGA Institute Terms and Conditions

8. Supplementary Figure 2
FLC tumors overexpressed neuroendocrine markers and their integration in the FLC molecular classification. (A) Plot representation of neuroendocrine gene marker expression in FLC patients compared with nontumoral samples by fold-change. FLC tumors present a significantly higher expression of Versican. Vertical axis: gene expression of samples; horizontal axis; FLC tissues (left) vs nontumoral tissues (right). (B) Integration of the neuroendocrine gene expression of FLC patients and nontumoral tissues in the molecular classification, and their enrichment, when significant, in the FLC classes.
Gastroenterology  , e10DOI: ( /j.gastro )
Copyright © 2015 AGA Institute Terms and Conditions

9. Supplementary Figure 3
Heat-map representation of FLC chromosomal alterations and scheme of the DNAJB1–PRKACA fusion transcript. (A) Heat-map representation of FLC chromosomal alterations. Each tumor is shown in separate columns, grouped by molecular class, and chromosome positions are indicated along the y-axis. Chromosomal amplifications and deletions are shown in red and blue, respectively. (B) Scheme of the DNAJB1–PRKACA fusion transcript and example of the Sanger sequencing result of the reverse-transcription polymerase chain reaction product from a FLC sample with the fusion transcript in the intersection point between the end of exon 1 of DNAJB1 and the start of exon 2 of PRKACA.
Gastroenterology  , e10DOI: ( /j.gastro )
Copyright © 2015 AGA Institute Terms and Conditions

10. Supplementary Figure 4
Somatic mutations found by WES and technical validation by TES and Sanger sequencing. (A) Distribution of the mutations obtained using SIFT and PolyPhen2 algorithms as noncoding and coding, synonymous and nonsynonymous, missense and nonsense, and, finally, among damaging, probably or possibly damaging, and benign. (B) Alignment visualization of TES results through Golden Helix Genome Browse software. (C) Results of the technical validation of the damaging mutations found in BRCA2 by Sanger sequencing.
Gastroenterology  , e10DOI: ( /j.gastro )
Copyright © 2015 AGA Institute Terms and Conditions