Matt:LabNotes/2014-11-1

From ZhangLabWiki
Jump to navigation Jump to search

Design FISSEQ Padlock Probeset for in situ cDNA Capture (New Genelist)

Get Genes

Build Target File for ppDesigner

Biomart: Get Exon Locations for Genes on Reference Genome

  • Use 409 HUGO names as input
  • GRCh38/hg38 reference genome
  • Unfortunately when try to output 'Associated Gene Name' (aka HUGO Official Gene Symbol) Biomart gave error so had to output Ensembl Gene IDs
    • Some LRG (locus reference genome) gene names and HSCHR (haplotype variants) outputted so will filter those out
  • Output parameters:
    • Ensembl Gene ID
    • Ensembl Transcript ID
    • Chromosome
    • Exon Rank
    • Exon Start (bp)
    • Exon End (bp)
    • Strand

Create Hash to convert Ensembl to HUGO

  • Using media:TopDiffExpGenes_1000hNucDataset_10282014.xlsx created hash table for converting Ensembl IDs to HUGO since Biomart output had to be ENSG
    • Hash naturally removes duplicates but also checked to see if there was gene names that didn't match 1-to-1
      • ENSG000000236922 matched to LINC01378 and AC092661.1
      • ENSG000000228918 matched to LINC01344 and GS1-122H1.2

Create Target File v2: Target consitutive exons (no crossing exon boundaries)

  • Changed the algorithm to only select intervals that don't contain boundaries of exons (like constitutive exons but doesn't have to be present in all transcripts)
    • Ordered all boundaries from least to great regardless if start or stop
      • In another array of equal size kept track of whether it was start or stop boundary
    • Selected intervals that on the left have a start boundary and on the right has a stop boundary with no boundaries in the middle
  • CreateTargetFile_v2.pl
    • input: biomart_export_newGenes.txt
    • output: target_file_newGenes.txt
  • Tested with one gene and two genes consisting of multiple overlapping exons
    • Runs with warnings but behaves correctly
    • Hash for converting Ensembl to HUGO missed a few genes due to multiple Ensembl Gene IDs for same HUGO IDs so had to do them manually
      • KCNMB2
      • NBPF15
      • NEFL
  • sortTargetFilesByChr.pl
    • Removes 25bp from each end of target
    • Switch strand from + to - and - to +
    • Also runs with warnings that 'given' and 'when' are experimental

Run ppDesigner

 ppDesignerCommands.sh

Consolidate OutputFiles and Convert to 0gap

  • Concatenate output files into 1 file: outputFile_1gap.txt
 cat outputFile_chr*.txt > outputFile_1gap.txt

Filter Out Bad Probes

  • Align probes to reference genome and mRNA
 perl  Probes2fasta.pl < outputFile_0gap.txt > outputFile_0gap.fa
 /home/kunzhang/softwares/Novocraft/novocraft/novoalign -d /home/kunzhang/RNAseq/Data/CommonFiles/refMrna.ndx -f outputFile_0gap.fa -F FA -r ALL > outputFile_0gap_novoalign_refMrna.out &
 /home/kunzhang/softwares/Novocraft/novocraft/novoalign -d /home/kunzhang/softwares/Novocraft/novocraft/human_g1k_v37 -f outputFile_0gap.fa -F FA -r ALL > outputFile_0gap_novoalign_human_g1k_v37.out &
 
  • Filter out probes that map to multiple locations in genome or don't map to refMrna
 perl  CleanupProbelist.pl
  • 3,553 probes
    • 389 genes
    • 2,813 exons

Filter Probes with Low Coverage based on Brain RNA-seq Data from Rui/Blue

  • Must still filter out probes that target regions with low coverage based on RNA-seq data
    • Sequencing data for transcripts from tissue Brodmann Area 8,10,17,21,22,41
 /media/LTS_33T/RL_LTS33T/201404_201405_7Samples_BulkNucleiBatch1-20140623_Expt146/STAR/
 samtools merge ~/scratch/FISSEQ_cDNA_ProbeDesign/BrainBulkTissueRNA_mappedhg19.bam RL-BA10-t-N707-15May14_S7_mapped/RL-BA10-t-N707-15May14_S7_Aligned.sorted.bam RL-BA17-t-N703-15May14_S3_mapped/RL-BA17-t-N703-15May14_S3_Aligned.sorted.bam RL-BA21-t-N705-15May14_S5_mapped/RL-BA21-t-N705-15May14_S5_Aligned.sorted.bam RL-BA22-t-N709-15May14_S2_mapped/RL-BA22-t-N709-15May14_S2_Aligned.sorted.bam RL-BA41-t-N711-15May14_S4_mapped/RL-BA41-t-N711-15May14_S4_Aligned.sorted.bam RL-BA8-sec9-t-N701-15May14_S1_mapped/RL-BA8-sec9-t-N701-15May14_S1_Aligned_chr12_16.sorted.bam
 samtools sort BrainBulkTissueRNA_mappedhg19.bam BrainBulkTissueRNA_mappedhg19.sorted
 samtools index BrainBulkTissueRNA_mappedhg19.sorted.bam BrainBulkTissueRNA_mappedhg19.sorted.bam.bai
  • Convert probe outputFile_0gap_filtered.txt (from CleanupProbelist.pl) to BED file: ConvertProbesHg38_2_BedHg19.pl
    • Also converts hg38 positions to hg19 to match RNA-seq data
      • Uses position alignments to human_g1k_v37 to get hg19 coordinates
      • Checked new coordinates with UCSC Genome Browser for genes on forward and reverse strands - both were correct
 coverageBed -abam BrainBulkTissueRNA_mappedhg19.sorted.bam -b outputFile_0gap_hg19.bed > outputFile_BrainBulkTissueRNA.cov
  • Rank Order plot of Probes and the number of reads that overlap with each probe target region
    • 523 of 3,933 probes had 0 overlap

File:RankOrderProbesOverlapReads.JPG

  • Filter out probes with <10 reads overlapping target region
  FilterProbelist_LowReadCounts.pl

RevComp Probes for potential RNA capture

  • RevComp() each arm and then swap arms (columns 5 and 7)
  • RevComp() target region (column 11)