Matt:LabNotes/exonProbeDesign

From ZhangLabWiki
Revision as of 21:08, 7 December 2012 by >Mzcai
Jump to navigation Jump to search

Finding constitutive exon target regions from gene list

1) Start with a list of genes of interest with Entrez IDs and Official Gene Symbols

2) Use Ensembl's Biomart on Entrez ID list and output to TSV file Settings:

  • Dataset
    • Homo sapiens genes (GRCh37.p8)
  • Filters
    • EntrezGene ID(s) [e.g. 100287163]: [ID-list specified]
    • Status (transcript) : KNOWN
  • Attributes
    • Ensembl Gene ID
    • Ensembl Transcript ID
    • Exon Chr Start (bp)
    • Exon Chr End (bp)
    • Ensembl Exon ID
    • Exon Rank in Transcript
    • Constitutive Exon
    • Associated Gene Name
    • Strand
    • Gene Start (bp)
    • Gene End (bp)
    • Transcript count
    • Chromosome Name

3) Insert Entrez IDs into column 13 of Biomart output

  • may need to use Biomart again with Attributes: Ensembl Gene ID, Ensembl Transcript ID, and EntrezGene ID and then Excel to VLOOKUP Entrez with Gene ID

4) Use original gene list and edited Biomart output to run ConstitutiveExonFinder.pl

(If chromosome and strand info available, they are used for checking that genes from the two input files match. If gene start/stop info available, they are used to set boundaries of target region. Only Symbol and Entrez ID are required to run and all other columns can be set to NA and run with ConstitutiveExonFinder_entrezOnly.pl.)

  • Finds the 5 regions with the most number of transcript overlaps
  • Conditions for finding regions:
    • Region must be greater than $minLength (default 100bp), which can be set in first line of code
    • Region must not have save start/end index as a previously found region (to prevent picking the same exon 5 times)

5) Remove Blank Lines and Duplicates

  • Something about Biomart can create duplicates and blank lines sometimes, run RemoveBlankLinesandDuplicates.pl to clean up file if needed
  • If gene is duplicated with one corresponding to normal chromosome and the other to an assembly patch, assembly patch regions are removed
  • For assembly patches that aren't duplicates, they are changed to chromosome number (e.g. HSCHR5_1_CTG1 is changed to Chr5)

6) Find missing genes in target file

  • Biomart sometimes misses genes from input
  • Run FindMissingGenes.pl to identify missing genes and repeat biomart and ConstitutiveExonFinder.pl for those genes and add onto end of target file

7) Sort target regions into separate files by chromosome number and remove 25 bp from each end with sortTargetFilesByChr.pl

  • MUST FIRST:
  1. manually remove columns: Entrez ID, Overlap, and Transcript Ratio and move Strand column to end
  2. Turn ExonEnd values from negative to positive values
  3. Add 'chr' in front of Chromsome number (e.g. 18 -> chr18)
  • Final column order should look like: GeneSymbol_# Chr ExonStart ExonEnd Strand
  • Remove 25bp from each end because 25bp is the max length of probe arm and this avoid the probe annealing to a splice junction
  • These are the target files for ppDesigner

Running ppDesigner

Parameters:

probe_parameters.pl

  • $primerMaxTm = 65
  • $primerMinTm = 55
  • $flankingLn = 25

jobFile.pl

  • $primerMaxLen = 25
  • $primerMinLen = 20
  • $H1_plus_H2_Len = 50
  • $H2OptLen = 22
  • $H1OptLen = 22

If gap size is 20

  • $targetMinLen = 20
  • $targetMaxLen = 20

If gap size is 0

  • $targetMinLen = 1
  • $targetMaxLen = 1

Generated Probes Summary

Original 5k genes

For the 86 stem genes that were missing from the original probe results

  1. got cDNA sequences from biomart
  2. designed probes for these transcripts (903 total, some genes had as many as 39 transcripts, average was 10)
  • Note: These two gene names correspond to same gene, there may be others
    • PPPDE2 -> DESI1
    • SIP1 -> GEMIN2
  • if multiple transcripts, designed probes for each one naming them v1, v2, etc

Result: ~2,800 0gap probes and ~3,000 20gap probes