(N) Expression data from RNA-seq (medium/slow, needs A)

[mrshu]

• The goal is to display the results of measurement of expression (amount of mRNA) by RNA-seq
• Workflow:
  • The original data are reads in fastq format. Some preprocessing can be done (quality trimming etc)
  • Reads are aligned to the genome to produce sam/bam file. This is SLOW. The file is then sorted and indexed.
  • Bam files can be used in the browser, but they are big. We will report only the number of reads at each position in a wig (wiggle) format.
  • Wig files can be loaded to the database but perhaps more efficiently converted to binary bigwig files. The database then contains only reference to bigwig file.
• Data:
  • malGlo [8] - only reads provided. Out of 27 experiments choose only 1-2, align to genome, e.g. this one: [9]
  • malSym [10] - bam files provided
• malGlo needs to align reads to the genome.
  • Currently recommended aligner is STAR https://github.com/alexdobin/STAR
  • It seems that STAR can directly create wig files, read the manual for recommended settings (e.g. the section on small genomes)
  • To convert wig to bigwig, use wigToBigWig on genomika
  • To load bigwig file, see commands below
• malSym already has bam files for several experiments
  • These need to be converted to wig / bigwig
  • First use bedtools suite to create bedgraph (see commands below), then convert to bigwig using bedGraphToBigWig (installed on genomika)
  • To load bigwig file, see commands below
  • Multiple experiments are better combined to a single composite track with individual subtracks
  • Subtracks are loaded to db normally, composite tract is noted only in trackdb file, see below
• Useful commands (modify for your situation):

  
  # bam to bedgraph 
  faSize -detailed genome.fa > genome.sizes
  bedtools genomecov -ibam reads.bam -g genome.size -bga -split > reads.bedgraph

to create track, place bigwig file to appropriate place in /gbdb

then create table with reference to this file:

hgsql malXyz1 -e "CREATE TABLE table_name (fileName varchar(255) not null);" hgsql malXyz1 -e "insert into table_name values ('/gbdb/malXyz1/filename.bw');"

in trackDb.ra include something like this: (change 500 to appropriate value at which read depth is clipped)

track table_name shortLabel RNA-seq coverage longLabel RNA-seq coverage visibility dense group rna type bigWig 0 500

composite track from multiple experiments:

track track_name compositeTrack on type bigWig 0 200 shortLabel RNA-seq coverage longLabel RNA-seq coverage group rna visibility dense

track subtrack_name shortLabel subtrack_label longLabel subtrack_label parent track_name type bigWig 0 250 visibility full maxHeightPixels 80:16:8

- Last year notes: https://github.com/fmfi-genomika/genomika-2017/wiki/Expression-tracks
  - However steps there are mostly not recommended this year
  - Last year tracks, see RNA-seq WT1 (L) in yarLip browser