bioinformática y patología moleculare-pat.org/dvd/pdf/edealava.pdf · patología molecular...

Bioinformática y

Patología Molecular

ENRIQUE DE ALAVA

Esquema

• Introducción: dos consideraciones

• Aplicación: Tres ejemplos

• Reflexión final

JANO



Diagnóstica

• LIMS

• Gestión de peticiones electrónicas

• Visión web de informes

• Integración en la historia clínica

• Análisis de imagen

• … como las otras unidades del Servicio

Bioinformática

La aplicación de tecnología

de computación a la

gestión y análisis de datos

biológicos (Wikipedia)

Agosto de 2011

Agosto de 1994

F. Collins

J. Craig Venter

15 de febrero de 2001

Esquema




Mujer de 68 años con dolor abdominal súbito

Abril de 2001

Diseño de primers

Alineamiento E11

c-kit ACCTACAAAT ATTTACAGAA ACCCATGTAT GAAGTACAGT GGAAGGTTGT

E11 .......... ........AA ACCCATGTAT GAAGTACAGT GGAAGGTTGT

c3E11R CCTTTC...T CCCCACAGAA ACCCATGTAT GAAGTACAGC GGAAGGTTGT

c3E11F CCTTTC...T CCCCACAGAA ACCCATGTAT GAAGTACAGC GGAAGGTTGT

ALINEAMIENTO DE SECUENCIAS PROTEICAS

1q gain and CDT2 overexpression

underlie an aggressive and highly

proliferative form of Ewing Sarcoma

Carlos Mackintosh, José Luis Ordóñez, Daniel J. García, Victoria Sevillano, Antonio

Llombart-Bosch, Karoly Szuhai, Katia Scotlandi, Marco Alberghini, Raf Sciot, Friedl Sinnaeve, Pancras C.W. Hogendoorn, Piero

Picci, Sakari Knuutila, Uta Dirksen, Maria Debiec-Rychter, Karl-Ludwig Schaefer,

and Enrique de Álava

Oncogene 2011

Unpaired integrative analysis

Functional validation • CNA in EWS cases (EE99) - set 1 (n=67)

studied with arrayCGH and correlation with clinical data.

• Expression microarrays in EWS cases – set 2 (n=38).

– Unsupervised and supervised analysis in groups made along with their 1q - genomic status.

• Functional validation of a candidate gene located in 1q

BAC-CGH array: methodology

aCGH:

Frequency of CNAs by chromosome

aCGH Analysis

• Data acquisition: GenePix software

– Flagging with GPix scripts (removal of bad quality spots)

• Data processing: Segmentation, Gain/Loss Calling.

– Packages snapCGH, CGHcall (BioConductor)

• Reduction of complexity/Clustering: CGHregions, WECCA (BioC)

• Correlation with clinical data:

– CGHtest, CGHPermutations and CGHMultiarray (BioC)

• Evaluation of the overall size of genome altered

– Script “CGH altered genome”, written by myself in R language

Bioinformatics screening of CNA’s

clinical impact Test type: CGHPermutations

Test CNAs found enriched p-value / FDR q-value

#1.- Primary tumors with osseous location vs. non-

osseous NS

#2.- Primary tumors with axial location vs. those

with peripheral location NS

#3.- Relapsing tumors (metastasis and/or local

relapse) vs. non–relapsing

1q gain (proximal)

1q gain (distal)

0.015 / 0.048

0.009 / 0.048

#4.- Metastasic primary tumors vs. non-metastasic NS

#5.- Metastasis at diagnosis vs. new metastasis NS

#6.- Single metastasis vs. multiple metastases NS

#7.- Metastasis to sites including lung vs.

metastasis to sites not including lung

8p gain (entire arm)

8q gain (entire arm)

0.017 / 0.03

0.002 / 0.005

#8.- Patients younger than median age vs. older 16p loss (pter – 2Mb) 0.03 / 0.057

#9.- Cell lines vs. tumors with PGA > 1% 8q gain (124Mb – 129 Mb from pter)

3p loss (40Mb – 75Mb from pter) 9p loss (pter – 24 Mb from pter)

0.003 / 0.06

0.004 / 0.02 0.005 / 0.014

#10.- 1qGT vs. 1qNT Chr 12 gain

Chr 20 gain

16q loss (entire arm)

0.005 / 0.005

0.002 / 0.002

0.02 / 0.07

Test type: CGHlogrank

CNAs found significant OS (p-value/FDR q-value) DFS (p-value/FDR q-value)

1q gain (proximal) < 10-3 / < 10-3 < 10-3 / < 10-3

1q gain (distal) 0.002 / 0.021 0.004 / 0.03

Chr 20 gain 0.006 / 0.026 NS

Chr 22 loss 0.014 / 0.048 NS

Chr 10 loss NS 0.005 / 0.06

Lenguaje R

Classification of the ES set 2 according to

the 1qG signature (1qGSig)

• Signature Definition:74 genes located in 1q with the highest d-values and fold-change values (cut-off settled: 90th percentile, d-value > 4.9; fold change, R-fold > 1.5)

• 9 tumors out of 38 clearly positioned in a same cluster (including the 6 tumors with known 1qG)

1qG Sig

Lancet 2011; 378: 1812–23

Classification of the ES set 2 according to

the 1qG signature (1qGSig)

• Signature Definition:74 genes located in 1q with the highest d-values and fold-change values (cut-off settled: 90th percentile, d-value > 4.9; fold change, R-fold > 1.5)

• 9 tumors out of 38 clearly positioned in a same cluster (including the 6 tumors with known 1qG)

1qG Sig

Prospective

clinical trial

Genomics

Proteomics

Functional validation

(in vitro/in vivo)

Tissue Validation xenografts

Cell lines

GE Animal

models

shRNA screens

Bioinformatics Biobanks

• MSigDB database C2: match with gene sets unveiling a lost of the cell

cycle controls

• MSigDB database C4: enrichment in hundreds of GO terms involved in

cell cycle regulation

• MSigDB database C3: enrichment in datasets belonging to genes with

promoters containing E2F - response elements

• MSigDB database C1: enrichment in genomic positional 1qG genes.

GSEA analysis: 1qGSig vs. non-

1qGSig tumors

GSEA genesets in 1qGSig tumors - IPA validation

C2 MSigDB Size NES NOM p-val FDR q-val

G1 to S cell cycle reactome 68 2.193 0.000 0.001

Pyrimidine metabolism 57 1.989 0.000 0.005

P21 any dn 34 1.967 0.000 0.006

Ren e2f1 targets 38 1.945 0.000 0.006

Schumacher myc up 49 1.938 0.000 0.006


V$E2F Q4 01 185 2.128 0.000 0.002

V$E2F1DP1 01 185 2.009 0.002 0.002

V$E2F1 Q6 01 192 1.998 0.002 0.002

V$MYC Q2 152 1.799 0.000 0.010

V$MYCMAX 01 215 1.763 0.000 0.015


Cell cycle go 0007049 294 2.114 0.000 0.015

Regulation of cell cycle 170 2.108 0.000 0.006

Mitotic cell cycle 140 2.098 0.000 0.004

Dna replication 93 2.095 0.000 0.004

G1 S transition of mitotic cell cycle 27 1.971 0.002 0.006

Regulation of cyclin dependent protein kinase activity

42 1.942 0.000 0.008


CHR1Q44 42 2.333 0.000 0.000

CHR1Q42 102 2.187 0.000 0.006

CHR1Q25 73 2.022 0.002 0.034

CHR1Q22 67 1.960 0.002 0.053

CHR1Q21 212 1.915 0.002 0.057

CHR1Q41 36 1.779 0.005 0.105

CHR1Q24 48 1.720 0.014 0.147

CHR1Q32 143 1.666 0.014 0.210

CHR1Q23 78 1.630 0.014 0.235

SAM analysis to identify those 1q genes with

the highest change in expression

• ES set 2 SAM. D-value-ranked list

AFFY_ID SYMBOL CHR bp_position p.value d.value stdev q.value R.fold

222680_s_at DTL 1 210275541 0 11.8830232 0.18993904 0 4.78013054

204603_at EXO1 1 240078157 0 10.7803037 0.20495983 0 4.62524311

206102_at GINS1 20 25336322 0 10.4319029 0.1801058 0 3.67780813

208644_at PARP1 1 224615014 0 10.3203701 0.10240054 0 2.08032868

200750_s_at RAN 12 129922520 0 10.1432607 0.10119041 0 2.03693604

202420_s_at DHX9 1 181075073 0 9.65949114 0.07362746 0 1.63716957

210415_s_at ODF2 9

c(130258252,

130259107) 0 9.52624618 0.08992072 0 1.81077454

209825_s_at UCK2 1 164063513 0 9.51261952 0.15449537 0 2.76955779

211594_s_at MRPL9 1 149998746 0 9.21492519 0.0871298 0 1.74459795

207828_s_at CENPF 1 212843154 0 9.12066268 0.21432169 0 3.87650316

211519_s_at KIF2C 1 44978076 0 9.0995555 0.18756768 0 3.26431996

223229_at UBE2T 1 200567407 0 8.95465463 0.20409681 0 3.54950146

203316_s_at SNRPE 1 202097362 0 8.95076944 0.16118798 0 2.71839749

207332_s_at TFRC 3

c(-197260551, -

197260551) 0 8.9270326 0.22422384 0 4.00458714

First term of the list: high and consistent change

Fold change markedly higher than the expected due to

genomic dosage (1.5)


Functional validation • CNA in EWS cases - set 1 (n=67) and 16

EWS cell lines studied with arrayCGH and correlation with clinical data.




1qG cell

lines are

dependent in

vivo on

CDT2 for

proliferation

DTL belongs to the CUL4/DDB1 ubiquitin ligase complex.

It selects the targets to be tagged for degradation

-Cdt1

-P53

-P21 & p27

E2f1

Kim, Y., Starostina, N. G., and Kipreos, E. T. (2008). The CRL4Cdt2 ubiquitin

ligase targets the degradation of p21Cip1 to control replication licensing.

Genes Dev 22, 2507-2519.

Cell Cycle 5:15, 1719-1729, August 2006

GENES & DEVELOPMENT 22:2496–2506, Sept 2008

GENES & DEVELOPMENT 22:2507–2519, Sept 2008

THE JOURNAL OF BIOLOGICAL CHEMISTRY 283: 43,

29045–29052, August 2008

Developmental Cell 15, 890–900, December 2008

Cell Cycle 5:15, 1675-1680, August 2006

Molecular Cell 23, 709–721, September 1, 2006

GENES & DEVELOPMENT 20:3117–3129, Nov 2006

Ongoing work: preclinical validation in ES of a specific

inhibitor of the CRL protein-ubiquitin-ligase complexes

Meyerson et al., Nat Rev Gen 2010

XS Puente et al. Nature 000, 1-5 (2011) doi:10.1038/nature10113

Profile of somatic mutations in four CLL genomes.

XS Puente et al. Nature 000, 1-5 (2011) doi:10.1038/nature10113

Mutational and functional analysis of MYD88 in CLL.

Esquema




Aspectos relevantes

• Análisis de secuencias

• Anotación de genomas

• Análisis de la expresión génica

• Análisis de mutaciones en el cáncer

• Análisis de la expresión de proteínas

• Modelado de sistemas biológicos

• Análisis de imagen de alto rendimiento

Esquema




Esquema



• Reflexión final. Tiempo de cambio

BIOMEDICINA Y ASISTENCIA SANITARIA: CONCEPTO ACTUAL

ESTILO DE VIDA

SALUD ------------- ENFERMEDAD

AMBIENTE

DOTACIÓN GENÉTICA

RESISTENCIA ------------- SENSIBILIDAD

GENOMA PROTEOMA

MEDICINA PERSONALIZADA

TRATAMIENTO

Through the study of expression

profiles in routine samples

2500 5000 7500 10000 12500 15000 17500 20000

• Prepare matrix coated tissue section (on ITO slide)

• Obtain MALDI spectra from defined positions (pixels)

200 µm pixels

MALDI imaging

integrando proteómica e imagen

Momento de cambio

• Progreso tecnológico rápido con

producción masiva de datos

• Progreso en bioinformática-computación

• Biología de sistemas frente a

reduccionismo

• Acceso a la personalización (genoma)

• Globalización

¿Patología de sistemas?

Costa J. Arch Pathol Lab Med 2008

“The Anatomy Lecture of Dr. Nicolaes Tulp” – Rembrandt, 1632

After “The Anatomy Lecture of Dr. Nicolaes Tulp” – Rembrandt, 1632

(Cortesía Dr. Carlos Cordón Cardó)

Gracias

[email protected]

bioinformática y patología moleculare-pat.org/dvd/pdf/edealava.pdf · patología molecular...

Documents