Bioinformatics tools and techniques Into the heart of darkness

1
Bioinformatics tools and techniquesInto the
heart of darkness

• Elaine Kenny
• Colm ODushlaine
• 15/11/07

2
Summary

• Simple overviews of some of the tools and methods
  used by EK and COD
• TK notebook
• get_hapmap_snps.pl retrieve HM genotype
  information for a list of SNPs
• GeneViewer.pl cross_ref.pl visualise e.g. SNPs
  in the context of other genomic landmarks. Score
  SNPs depending on how many of these landmarks
  they overlap with
• ld_expander.pl find SNPs in LD with SNPs of
  interest, based on user-specified r2 and LD
  window (distance between SNPs)
• STATA
• VIM command line text editor
• Lab website

3
TK notebook

• Application for saving notes, to-do lists, daily
  logs, and any other kind of textual information
  in a place where you can find it all again, and
  where related information is easily found
• Easy to edit and rapidly searchable
• DEMO editing
• DEMO search

4
get_hapmap_snps.pl

• Simple script to read in a 1-column list of SNPs
  and retrieve HapMap genotypes
• Can select population and strand
• DEMO
• Retrieved data can be loaded into HaploView
• DEMO

5
cross_ref_scored.pl

• Score SNPs based on how many putatively
  functional regions they overlap with
• On a per gene / chromosome basis
• Gene basis
• Type perl cross_ref_scored.pl file_A file_B
  file_C ...
• where
• file_A - 2-column file of SNPs (format
  id, location)
• file_B - 3-column file of EXONS (format
  id/name, start, stop)
• file_C ... - whatever you want, (format
  id/name, start, stop)
• i.e. other regions like CpGs,
  TFBS, clusters. Any order.

6
cross_ref_scored.pl example output
Can then be merged with HapMap / Perlegen to
retrieve MAF data for SNPs
7
Merge cross_ref_scored data with HapMap/ Perlegen
data using merge_per_hap.pl

• Type
• perl merge_per_hap.pl perlegen.txt hapmap.txt
  overlapped_region_scored.txt
• Where
• hapmap.txt 3-column file (format rsid,
  ref_allele, ref_allele_freq),
• perlegen.txt 3-column file (format rsid,
  ref_allele, ref_allele_freq)

8
cross_ref.pl applied to WGA data

• cross_ref.pl Scoring SNPs throughout genome
• Data analysed on coding/non-coding basis
• (coding)
• perl cross_ref.pl Overlapped_regions_scored.WTCCC.
  chr22.coding.txt 22 WTCCC_T2D_chr22_without_inferr
  ed.forCrossRef WGA_databases/coding_non_synon_SNPs
  _UCSC.clean3 WGA_databases/coding_synon_SNPs_UCSC
  .clean2 WGA_databases/RefSeq_Genes_UCSC.byExon.u
  niqid1 WGA_databases/Triplexes_may2006.bed2
  WGA_databases/splice_site_SNPs_UCSC.clean2 gt
  Overlapped_regions_scored.WTCCC.chr22.coding.log
  
• (input-dependent, coding/non-coding dependent,
  arbitrary)
• (noncoding)
• perl cross_ref.pl Overlapped_regions_scored.WTCCC.
  chr22.NONcoding.txt 22 WTCCC_T2D_chr22_without_inf
  erred.forCrossRef WGA_databases/TFBS.chr221
  WGA_databases/CpG_islands_UCSC.uniqid1
  WGA_databases/Most_conserved_phastConsElements17wa
  y_UCSC.clean1 WGA_databases/promoters_knowngene_h
  g18.txt1 WGA_databases/sno_or_miRNA_UCSC.uniqid1
  gt Overlapped_regions_scored.WTCCC.chr22.NONcoding
  .log

9
cross_ref.pl

• cross_ref.pl output
• Load into STATA. If SNPs have e.g. association
  p-values, calculate adjusted p-value (R. Anney)
  as -log10P
  cross_ref_score

10
GeneViewer.pl

• GeneViewer.pl Visualise overlapping features
  (e.g. exons, SNPs etc.) along e.g. your gene of
  interest (html output)

11
ld_expander.pl

• Find proxies (SNPs in LD) for a list of SNPs
• User specifies the r2 and LD window
• Currently configured to obtain proxies from HM
  CEU
• Result is a list of additional proxy SNPs that
  have been obtained by LD expansion
• DEMO
• Note dont LD expand gt150000 SNPs, or HapMap
  will ban you! COD has an alternative version
  that uses local pre-computed pairwise LD SNP files

12
STATA

• Extremely powerful and flexible
• gt65k rows handled shock horror!
• Can write scripts to automate tasks, e.g. read in
  file, do analysis, save results
• When use GUI to run some commands, the commands
  are shown in the command window, so can save in a
  do file
• COD, EK and R. Anney strongly advocate this as a
  platform for both file manipulation and
  statistical analysis

13
STATA example using WTCCC data
Bipolar Disorder, Coronary Artery Disease,
Crohn's Disease, Hypertension, Rheumatoid
Arthritis, Type 1 Diabetes, Type 2 Diabetes
14
DATA FORMAT

• 3 folders
• Basic
• Each case collection against the pooled control
  groups 58C and UKBS
• Combined cases
• Combining other case collections as controls
• Combined controls
• Combining phenotypically relevant case
  collections
• (e.g. RA/T1D, autoimmune )
• Data are split by chromosome

15
Questions

• How do I get all of the chromosome data for my
  gene of interest into one file?
• How do I search easily all of the SNP information
  for my gene(s) of interest?
• Create a .do file for all manipulations that
  you want to carry out to the data
• DEMO
• Good starting resource http//www.ats.ucla.edu/st
  at/stata/

16
VIM

• Vi Improved. Mainly UNIX but cross-platform
  text editor (available for Windows).
• Full list of commands outside scope of this
  demonstration
• Very fast and efficient, esp. with search and
  replace functions on large datasets
• Regular expression pattern matching
• DEMO
• Integrates with Cygwin (www.cygwin.com very
  useful UNIX emulator for windows)

17
Group website

• Some useful stuff up there!
• Please send information about current projects
  etc. Good for our image as a group and minimal
  effort required on your part
• DEMO

18
Conclusions

• Small summary of some things you can do
• Slides and video demonstrations will be online
  at http//www.medicine.tcd.ie/psychiatry/research
  /neuropsychiatry/Protocols/
• COD EK available for advice (Fridays
  9-9.02am)
• These things will help you in your work!!