Evaluation of de novo transcriptome assemblies from RNA-Seq

For simplicity, we assume that RNA-Seq reads are sequenced uniformly across the transcriptome. In addition, we only consider fixed length single-end RNA-Seq reads. We denote the read length as L. Given a set of MT transcripts, we denote the relative expression levels as τ = (τ1, τ2, . . . , τMT ), the read generating probabilities as Θ = (θ0, θ1, . . . , θMT ) and the expected read coverage as Ξ = (ξ0, ξ1, ξ2, . . . , ξMT ). We further denote the lengths of transcripts as L = (l0, l1, . . . , lMT ) and let l0 = L. Here transcript 0 refers to a non-existing “noise” transcript. It represents all reads that are generated from the background noise.