When assembling transcriptomes based on a reference why are the 5' and 3' UTRs commonly lost in the data?

Question

[cxu325]

Compared to genome assembly, why are the UTRs lost when assembling transcriptomes based off a reference?

Is it because UTRs are not as tightly constrained so that they have less sequence conservation than ORFs. Therefore because they change more they don't align as well to the reference? Thanks!

[Amitm]

hi, I dont have much experience in de novo transcriptome assembly but I have noted that many times the coverage of a given transcript is biased towards the 3' end. The reason I read was that because decay starts from 5' end and hence comparatively more reads support towards the 3'. So I guess because of some amount of decay (at 5' end) and highly variable regions (mostly at 3' end), the transcript start-end points will be difficult to assemble for the algorithm, as compared to internal constitutive exons.