heterogeneity Population > patient > tissue > genome - PowerPoint PPT Presentation

Dissecting cancer heterogeneity Population > patient > tissue > genome Florian Markowetz CRUK Cambridge Institute www.markowetzlab.org

Heterogeneity in cancer                     Inter-patient Intra-patient Intra-tumor Intra-tumor population spatial, tissue genetic subtypes temporal

Systems Genetics of Cancer • What are prognostic subtypes of cancer? • Which genetic events drive tumour development? • What are markers to predict disease progression?

Population heterogeneity                     Curtis et al , Nature 2012

METABRIC – genomic and transcriptional landscape of breast cancer ~400 paired normals Dataset 2 mRNA ~1000 samples Copy number changes miRNA Dataset 1 SNPs ~1000 samples Histopathology Clinical information https://www.ebi.ac.uk/ega/studies/ EGAS00000000083

Intra-patient heterogeneity                     Schwarz et al , submitted Spatial and temporal heterogeneity in ovarian cancer predicts survival

Intra-patient heterogeneity in HGSOC HGSOC OV03/04 study • Multiple metastases • 17 patients • Good initial response • 3-30 samples per patient • Often resistant relapse • Biopsy, surgery and relapse • Genomic rearrangements • Pre- and post-chemotherapy

the genome a disease of Cancer is http://www.nasa.gov/images/content/514467main_41s_factoring_DNA_1024.jpg

the tissue a disease of Cancer is http://clincancerres.aacrjournals.org/content/18/16/4266/F3.expansion.html

Hanahan and Weinberg (2001)

Comprehensive portraits of cancer Genomics Tissue

Tumors are complex tissues Van ’ t Veer et al (2002) http://ms.lbl.gov Ross-Innes et al (2012) DNA RNA Protein ChIP

Intra-tumor heterogeneity                     Yuan et al, Science Trans Med 2012 Quantitative image analysis of cellular heterogeneity complements genomics

Automated image analysis Yinyin Yuan H&E Supervised classification Spatial smoothing Cell types and location

CRImage

Man vs Machine Raza Ali (Caldas lab)

Quantitative analysis of tumour composition

Spatial features of tissue organisation Uniform Clustered Spatial statistics (K-score)

Spatial features of tissue organisation

Rimm D L Sci Transl Med 2011;3:108fs8-108fs8

Spatial features of tumour tissue

Morphological heterogeneity Yinyin Yuan A C ancer Morphological H&E Median P rognosis features kewness Lymphocyte Genomic aberrations S S tromal S ignaling pathways S tandard Deviation Morphological features 1. Fraction of pixels outside of the circle with radius effr 2. Shape factor, 3. 1st Hus translation/scale/rotation invariant moment 4. Eccentricity calculated based on geometric information 5. Eccentricity calculated based on image moments

Morpho-genomic subtypes Yinyin Yuan

Morphology <-> Gene expression TOP1MT Yinyin Yuan FAM91A1 CMAS GINS4 MCM4 sd_C-g.I2 CASP2 PHF20L1 SLMO2 MASTL C19orf2 sd_C-g.ecc MTERFD1 RECQL4 HINT3 TUBG1 CSE1L sd_C-m.ecc YWHAZ UTP23 median_C-g.I1 BOP1 FOXM1 GMPS ATAD2 MCM10 POP1 RAD51AP1 median_C-g.acirc PDSS1 CCNE2 MTBP DSCC1 HAGH NCAPD2 skewness_C-g.ecc HSF1 EXOSC4

JAM3 – driver of cell morphology Yinyin Yuan Chris Bakal

Xin Wang

Stainings in Comparison to Anne Trinh tissue microarrays gene expression classifier

Spatial features are predictive Anne Trinh

ASUMT : A Still Unnamed MATLAB Toolbox Anne Trinh Feedback Box: Image * Instructions Properties * Errors Load & Preprocess: * Load TMA View * Detect Outline Results * Find Brown Area Regional Segmentation: * Label Dataset (file or directly) * KMeans-MRF (grayscale & RGB) Plotting Cell Count: Area * Background Thresholding * H-minima Watershed * SVM Cell Classification Save: * mat file * csv & images * classifiers

ER+ ERBB2 ampl HER2 expr IFISH = IF + FISH

Go IFISH: a toolbox for semi-automated detection of nuclei, membrane and spots Anne Trinh

Single cell analysis of stain intensities Anne Trinh HER 2 E R ERB B2

Key collaboration partners • Carlos Caldas, Raza Ali, Suet-Feung Chin, Oscar Rueda, Stefan Gräf @ University of Cambridge • Yinyin Yuan + lab @ Institute for Cancer Research • Chris Bakal + lab @ Institute for Cancer Research • JP Medema , Louis Vermeulen @ Amsterdam Medical Center • Anne-Lise Børresen-Dale , Hege Russnes, Inga Hansine Rye @ Oslo University

Alumni: the team Xin Wang Yinyin Yuan Roland Schwarz Mauro Castro Gökmen Altay

CRUK Cambridge Institute Carlos Caldas Paul Pharoah • Breast Cancer Strangeways Functional Laboratories, Genomics Cambridge • Cambridge • Genetic Breast Cancer epidemiology Research Unit Jason FM lab Carroll • ER biology • ChIP-seq in tumors Stephen Friend Sage Bionetworks Doug Fearon • Tumor immunology • Tumor microenvironment

Dissecting cancer heterogeneity Thank you ! Florian Markowetz CRUK Cambridge Institute www.markowetzlab.org

Systems Genetics = genome × phenotypes × conditions