•Alignment to genome
–Computationally expensive
–It is never a good idea to simply align RNA-seq data to the genome Need a spliced
aligner or a surrogate (such as including exon-exon junction sequences in ‘genome’)
•Alignment to transcriptome
–Reads deriving from non-genic structures may be ‘forcibly’ (and
erroneously) aligned to genes
•Incorrect gene expression values
•False positive SNVs
•Many other potential problems
•Assembly
–Low expression = difficult/impossible to assemble
–Misassemblies/fragmented contigs due to repeats
–Requires vast amounts of memory
Drawbacks for
each strategy
Transcript
model
Genome
Exon
junction