Simulation of transcriptome assembly for gene expression analysis
The main script is transcriptome-simulation.sh that calls
sample-fasta.R to sample 100 sequences from the file ena.fasta.
The script then simulates reads and runs a standard velvet-oases
transcriptome assembly pipeline, both with and without digital normalization.
Output files are
sim-assembly-results-{suffix for each assembly}.txt
which gives the number of known transcripts BLASTed by assembled transcripts and
sim-assembly-results2-{suffix for each assembly}.txt
which gives the mean length of known transcripts mapped, and the number of base pairs in the assembled transcripts to the number bp in the known transcripts (roughly the number of isoforms with 1=one isoform)
The script expression-simulation.sh continues the above through mapping reads to
the known and assembled transcripts to evaluate gene expression versus the simulated
expression levels from rlsim.
Programs required: rlsim simNGS trim_galore and dependencies khmer screed velvet oases R and some Bioconductor packages cdhit