lncRNAs are known to be mostly polyadenylated just like mRNAs. Thus, polyA+ selected RNA-seq library preparations (such as Illlumnia TruSeq) are able to capture lncRNAs. However, the human lncRNA MALAT1 has been shown to have alternative 3' end processing, resulting in a mature transcript without a canonical polyA tail. Rather, MALAT1 is stabilized by a triple helix structure.
I'm curious if anyone has expertise on this question: How are polyA+ selected RNA-seq libraries able to sequence MALAT1? For example, GTEx RNA-seq uses TruSeq polyA+ selection, but is able to quantify MALAT1 expression across all tissues, and MALAT1 expression is high. (You can see MALAT1 expression on GTEx's IGV viewer here).
MALAT1 does have a short internal polyA signal around chr11:65,506,095-65,506,116 (~20 nt) which is a constitutive component of the 3' triple helix structure. Is it plausible that the oligo-dT beads used in TruSeq select for this signal, even though it is short and not at the very end of the 3' end of the transcript? From what I understand, mRNA polyA tails are usually ~80 - 250 nt.
My expertise is more computational rather than experimental so any insight here is appreciated.
