department of pathology uc davis school of medicine jeff gregg, m.d. the development of an...

25
Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Upload: lindsay-strickland

Post on 16-Jan-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Department of Pathology

UC Davis School of Medicine

Jeff Gregg, M.D.

The Development of an Informatics Platform for the Characterization of Clinical Samples

Page 2: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples
Page 3: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Ovarian Cancer Project

• Develop correlative molecular fingerprints for better diagnosis, prediction of outcome, and, ultimately, early biomarker development– Gene expression profiles– Genomic profiles (e.g. gains and losses of

DNA)

Page 4: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Ovarian Cancer Project• leading cause of death from gynecologic malignancy in the United

States • fourth leading cause of cancer deaths in women in the United States • estimated 26,700 new cases of ovarian cancer were reported in 1996,

and approximately 15,000 women will lose their lives to advanced disease

• lack of effective screening strategies and the paucity of symptoms and signs in early stage disease

• 70% of women with ovarian cancer present with advanced stage disease

• Despite surgical debulking and conventional chemotherapy, long term survival for advanced stage ovarian carcinoma rarely exceeds 15-30%

Page 5: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Ovarian Cancer Project

• Prognostic factors– Patient age

• <50 years old, 5-year survival of 40%

• >50 years old, 5-year survival of 15%

– Tumor stage

– Tumor grade

– Residual disease

– Second-Look status

– Performance status

Page 6: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Ovarian Cancer Project

Tumor Type # Goal # CurrentSerous Adeno Ca 20 12

Serous borderline 20 8

Mucinous Adeno Ca 20 6

Mucinous borderline 10 4

Endometrioid Adeno Ca

10 6

Normal ovary 20 83

Page 7: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Figure 3. Twelve ovarian RNA samples were run on the 2100 Bioanalyzer. Graphs of fluorescence readings for each sample are shown on the right; these readings are translated into a more traditional ‘gel image’ on the right.

Ovarian Cancer Project

Page 8: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Ovarian Cancer Project

Normal

BorderlineMucinous

Poorly diff.Adeno Ca

Page 9: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Ovarian Cancer Project• genes upregulatedin the tumors

Page 10: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Ovarian Cancer Project• set of genes upregulated in Borderline tumors

Page 11: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Tissue Database Construction- Phase 1

• Easily implemented in multiple labs

• Usable for a diverse group of individuals– Computer illiterate to computer savvy

• Basic enough that several individuals could troubleshoot and add functionality

• Linked over multiple networks

• Relatively inexpensive

Page 12: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples
Page 13: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples
Page 14: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples
Page 15: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples
Page 16: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples
Page 17: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Projects

• Develop a web-based tissue request system– Track requests– Form reports of distribution of tissue– Send out automated emails to ask users about

quality of material, grants obtained with material, manuscripts published using the the material

– Billing Module

Page 18: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

The End Result…

A microarray image file A gene expression file

•After the chip is scanned, the probes on the array will “light up”

•Brighter probes = More cRNA molecules attached = Higher gene expression

Page 19: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Topics of Interest

I. The Microarray: An overview

II. Annotation

III. Data quality

Page 20: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

How does Affymetrix organize its data?

38433_at M76125 558 AXL receptor tyrosine kinase |enzyme|signal transducer|cell cycle…Human tyrosine kinase receptor (axl) mRNA

Probe set

Accession Number

Common name, functions, description, etc.

Page 21: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Project: Re-Annotation

• Every week, NCBI releases re-annotated human genomic data in a large text file

• PROJECT: Write a program that takes in an old gene information Excel file, scans the NCBI file for updated info, and properly updates the Excel file.

• Catch: The NCBI file is 233 MB of text (it can’t be opened in a word processor)

Page 22: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Commercial Re-annotation Programs

Page 23: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Commercial Re-annotation Programs

Silicon Genetics’ GeneSpring

The Dragon Database

Affymetrix’sNetAffx

-Failed to annotate all genes

-Left out plenty of info on annotated genes

-Failed to annotate all genes

-Only 500 genes at a time

-User-unfriendly

-Cannot save info as text file

Page 24: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Projects

• Develop an automated reannotation program that would take sequence information from the Affymetrix arrays, blast, reannotate with name, Unigene number, map position…..

Page 25: Department of Pathology UC Davis School of Medicine Jeff Gregg, M.D. The Development of an Informatics Platform for the Characterization of Clinical Samples

Projects

Goal: To Establish a ColdFusion interface to interact with 2 in-house ACCESS Databases as well as a, yet undeveloped Database based in Access, SQL, Oracle language.

Duties: Write code for Web interface using CF markup language. Ability to consult with project leaders to determine the best way in which to implement design. Link existing databases using SQL language.

Skills Required: Advanced HTML skills. Knowledge of CFML and SQL language.

Knowledge of Javascript and XML, a plus. Understanding of Web Server design, esp. Microsoft IIs or iPlanet Web Server 6.0. Advanced understanding of, and ability to implement Web Security protocols and equipment.