Preterm Birth Network: Knowledge Base and Field Synopsis

1
Preterm Birth Network Knowledge Base and Field
Synopsis
Developing and Updating Cumulative Evidence on
Genetic Associations Experiences to date

• Siobhan Dolan, MD, MPH
• Albert Einstein College of Medicine, Bronx, NY
• Network of Networks Meeting, November 2006,
  Venice, Italy

2
Pathological Uterine Distention
Inflammation
Activation of Maternal/Fetal HPA Axis
Decidual Hemorrhage Abruption
Infection - Chorion-Decidual - Systemic
Multifetal Preg Polyhydramnios Uterine
abnormalities

• Maternal-Fetal Stress
• Premature Onset of Physiologic Initiators

Prothrombin G20210A Factor V Leiden Protein C, S,
Z Type 1 Plasminogen MTHFR
Interleukins IL-1, IL-5, IL-8 TNF-a Fas L
Gap jct IL-8
PGE2 Oxytocin recep
CRH E1-E3
Mechanical stretch
Chorion Decidua

CRH
CYP1A1 GSTT1
Susceptibility to Environmental toxins

MMPs
proteases
uterotonins
pPROM
Uterine Contractions
Cervical Change
PTB
Adapted from Lockwood CJ, Paediatr Perinat
Epidemiol 20011578 and Wang X, et al.
Paediatr Perinat Epidemiol 2001 15 63
3
Preterm Birth Network Update

• Efforts to secure funding for a large
  collaborative study have been unsuccessful to
  date.
• A modest number of researchers continue to be
  funded individually through MOD, NIH, etc.
• Active researchers (internationally) remain
  interested in collaborating
• PREBIC the Preterm Birth International
  Collaborative and its subcommittee PREGENIA the
  Preterm Birth and Genetics International Alliance
  remain active with joint projects and an annual
  meeting (last one held 4/06 in Geneva and next is
  scheduled for 9/07 in Italy).

4
Preterm Birth Network Joint Projects

• Published paper on Genetic epidemiologic studies
  of preterm birth Guidelines for research in the
  American Journal of Obstetrics and Gynecology

5
Guidelines for Research
6
Preterm Birth Network Joint Projects

• The paper outlined
• Phenotypic criteria
• Study Design Including minimum and optimal data
  set for genetic epidemiology studies into preterm
  birth
• Considerations in the selection of a control
  population
• Candidate gene selection
• The paper represents a first step in
  standardization across the field.

7
Invited Scientific Session at the ASHG meeting
in New Orleans in October, 2006

• SESSION 48 Preterm Birth as a Common Complex
  Disorder The Implications of Applied Genomic
  Research
• 400 PM   Introduction. C. A. Moore, Centers for
  Disease Control and Prevention, Atlanta , GA.
• 405 PM   Heritability in prematurity New
  insights into the genetics of preterm birth. K.
  Ward, Kapi'olani Hospital, Honolulu , HI .
• 430 PM   Genetics to genomics A framework for
  approaching preterm birth as a common complex
  disorder. S. Dolan, March of Dimes, and Albert
  Einstein College of Medicine, White Plains , NY .
• 455 PM   Meta-analysis and knowledge
  integration Implications for research on complex
  disease. J. P. Ioannidis, University of Ioannina
  School of Medicine , Ioannina , Greece .
• 520 PM   Implications of applied genomic
  research for public health What will it take to
  bring down the rate of preterm birth? M. Khoury,
  Centers for Disease Control and Prevention,
  Atlanta , GA.
• 540 PM   QA

8
Invited Scientific Session at the ASHG meeting
in New Orleans in October, 2006

• Generated interest amongst geneticists
• Stimulated ASHG to consider reaching out to SMFM
  (Society for Maternal Fetal Medicine) as a key
  partner

9
Perinatal Genomics Knowledge Base

• Major focus of effort in the past year
• Grant application to develop a perinatal genomics
  knowledge base was submitted in response to a
  Translational Bioinformatics RFP to the National
  Library of Medicine at NIH on October 1, 2005.
• Reviews in early 2006
• positive on importance, collaborators, and
  approach
• negative on bioinformatics capacity
• Resubmission in conjunction with the
  Bioinformatics Shared Resource at Albert Einstein
  College of Medicine in NYC just took place on
  November 1, 2006.

10
Context for Perinatal Genomics Knowledge Base
11

• Translational engine
• data processing
• secondary data analysis
• etc.

12
Perinatal Genomics Knowledge Base

• Review asked us to be explicit about
• What information were collecting
• From where?
• What are we doing with it?

13
Knowledge base is organized in three major
components

• The knowledge base will have three major
  components
• Investigators
• Studies/data collection platforms
• Literature (published and grey)

14
What are we collecting?

• For the registry of investigators, the following
  key data will be collected from each contributor
• Last name, First name, Middle initial
• Institution
• Address
• PI / Lab
• Web site
• Email
• Phone number (s)
• Statement of research interests
• Key words describing research interests

15
What are we collecting?

• For the registry of studies or data collection
  platforms, the following key data will be
  collected about each study or data collection
  platform that is contributed.
• Last name, First name, Middle initial of PI
• Name of study
• Grant (if applicable)
• Web site for study
• Abstract of study
• Type of study (cohort, case control, family
  based, etc)
• cases/controls or in cohort or parent-child
  trios
• Definitions of cases and controls
• Eligibility criteria including inclusion and
  exclusion criteria
• Methods for assessing phenotype and genotype
• Quality checks (if any) on phenotype and genotype
• Missing data
• Location of study
• Dates of enrollment
• Types of participants
• Age of participants
• Race of participants
• Geographic area of study

16
What are we collecting?

• For the compendium of published and grey
  literature, the following data will be extracted
  for the knowledge base from each item to be
  entered
• Last name, First name, Middle initial of first
  author
• Last name, First name, Middle initial of all
  subsequent authors
• Type of information (abstract, conference
  proceeding, unpublished data, published journal
  article)
• Peer-reviewed or not
• Citation information with link to PUBMED abstract
  or other electronic access if available
• For each study, the following data about gene
  disease association will be extracted and entered
  into the knowledge base
• Gene studied with links to Entrez gene at
  http//www.ncbi.nlm.nih.gov/entrez/query.fcgi?dbg
  ene
• Protein studied with links to Entrez protein at
  http//www.ncbi.nlm.nih.gov/entrez/query.fcgi?dbp
  roteincmdsearch
• Polymorphism studied with links to dbSNP at
  http//www.ncbi.nlm.nih.gov/SNP/
• Chromosomal location studied
• Keywords
• Outcome studied (preterm birth, very preterm
  birth)
• Modifiers (environmental agents)
• Ethnic group (Caucasian, continent, etc)
• Population studied (country)
• Type of study and cases / controls, triads,
  in cohort
• Phenotype description Spontaneous, Ruptured
  membranes, chorioamnionitis, exact gestational
  age in weeks
• Population genotyped maternal, fetal, paternal

17
What are we collecting?

• The data will be extracted in a standardized
  manner and will be stored using controlled
  vocabularies. This includes HUGO (Human Genome
  Organization nomenclature), ICD-10 (International
  Classification of Disease-10th revision), UMLS
  (Unified Medical Language System Metathesaurus)
  for genes and outcomes.
• The clinical data outlined in the Guidelines
  paper by Pennell et al. will be used as the
  starting point for the clinical and phenotypic
  vocabulary, and will be refined by the project
  team with input from the Advisory Committee.

18
Sources of Evidence

• The published literature
• PUBMED http//www.ncbi.nlm.nih.gov/entrez/query
  .fcgi?DBpubmed
• EMBASE http//www.embase.com/
• ISI Science Citation Index - http//scientific.th
  omson.com/products/sci/
• BIOSIS - http//www.biosis.org/
• PLoS (Public Library of Science) -
  http//www.plos.org/index.php
• The HuGE Published Literature database
  (http//www.cdc.gov/genomics/hugenet)
• The grey literature
• Conference proceedings
• Books
• Abstracts
• Technical reports
• Journals not available in electronic searches
• Conference proceedings and abstract books of
  highest yield will be those from the American
  Society of Human Genetics, Society for Maternal
  Fetal Medicine, Society for Gynecologic
  Investigation, American College of Medical
  Genetics, and the American College of
  Obstetricians and Gynecologists.

19
PubMed Search Strategy
20
EMBASE Search Strategy
21
EMBASE Search Strategy
22
EMBASE Search Strategy
23
EMBASE Search Strategy
24
HuGE Navigator
.
25
What are we doing with the data?

• Building and curating an online knowledge base
  that is publicly accessible.

26
What are we doing with the data?

• The relational database will have three major
  components that contain the data for
• the investigators (PKB_Contributor)
• studies/data collection platforms (PKB_Project)
• literature (PKB_Literature)
• Each will be linked by key fields including the
  investigators name and the linkages will be
  curated by the project team so that literature
  regarding various study collection platforms is
  appropriately linked and contributors are linked
  to the various study collection platforms, etc.

27
Preliminary Database Schema
28
What are we doing with the data?

• A fourth major component will be the gene
  annotation from the NCBI database which already
  resides on a server at the BISR.
• For each gene where a finding regarding
  prematurity has been entered into the knowledge
  base, the gene will be noted along with the
  chromosomal location, outcome (preterm, very
  preterm, etc), type of association, strength of
  the association, study details, etc.
• The knowledge base will use NCBI nomenclature for
  genes, proteins, and SNPs and will have alias
  mapping for users who might use non standard
  terminology such as that commonly used in
  obstetrics or by the clinical community.

29
Preliminary Database Schema
30
Alzheimer Research Forum as an example
31
Gene Overview
32
Opinions on Meta-Analysis ????
33
Our Progress to Date on Field Synopsis, by hand
Note Each row represents one published
report.
34
Table 1. Effect
35
Table 2. Study Details / Possible Biases
36
Table 3. Genotyping Concerns
37
Table 4. Interaction
38
Table 5. Biological evidence and comments
39
Preterm Birth Network Goals

• Attain funding to build the knowledge base
• Publish the high level field synopsis and perhaps
  a series of reviews
• Continue to look for large scale funding of a
  collaborative prospective project
• PREBIC annual meeting in September 2007

40
Thank you for your attention!
Special thanks to Muin Khoury John
Ioannidis Cynthia Moore Bruce Lin Marta Gwinn Wei
Yu