As sequencing costs have quickly dropped over the past decade, the reagents for RNA-seq library construction have become a larger share of the cost per experiment. So has the staff time required by the protocols, which may still take multiple days of bench work with plenty of opportunities for error and sample loss.
A new paper in Genome Research from our groups at Stanford and McGill describes a very streamlined protocol to solve that lingering problem and scale up gene-expression experiments to large sample sizes. The new method, Smart-3SEQ, uses the template-switching reverse transcription trick employed by Smart-seq and Smart-seq2 to generate double-stranded cDNA from tiny amounts of polyadenylated RNA, but with a twist: instead of reverse-transcribing every complete mRNA into a full-length cDNA and then fragmenting that down to sequencing size, Smart-3SEQ fragments the RNA first and uses template switching to add the sequencing adapters directly to the fragments. Then you can skip the usual Day 2 of library prep: all it takes from there is PCR and cleanup to get a sequencing-ready library.
Conceptual diagram of the Smart-3SEQ library preparation method
Hands-on steps are separated by horizontal lines A: Total RNA is denatured and fragmented by hydrolysis. B: The oligo(dT) primer, including a partial sequencing adapter, anneals at the beginning of the poly(A) tail. C: Reverse transcriptase synthesizes first-strand cDNA from the RNA template, and adds non-template dC at the end of the new strand. D: The second primer, which includes a second partial sequencing adapter, anneals to the new dC overhang. E: Reverse transcriptase synthesizes the second cDNA strand using the first as a template. F: After steps C–E, which occur consecutively in one incubation, the result is a double-stranded cDNA library with partial sequencing adapters at both ends. G: PCR with long primers amplifies the library and extends the adapters to full length, including multiplexing indexes. H: The only cleanup step in the protocol uses paramagnetic SPRI beads to purify the amplified library while excluding adapter dimers and short inserts. I: The final library contains the unknown cDNA sequence between the two sequencing adapters. The cDNA is sequenced in the orientation of the original RNA, yielding reads upstream of the end of the transcript.”
Because there are so few reaction steps, the protocol is very short and cheap with off-the-shelf reagents. And like other methods based on template-switching, it’s sensitive to very small amounts of RNA, even single human cells. However, there’s a price for prefragmenting your RNA: you only get a read from the 3´ fragment containing the beginning of the poly(A) tail, making this a 3´-end digital gene expression method (3SEQ), which can measure the expression level of a transcript but can’t distinguish alternate spliceforms. On the other hand, it brings another advantage: Smart-3SEQ also works on RNA that has already been fragmented, such as RNA from formalin-fixed tissue. In the next post , we’ll show a demonstration of this method in an extreme scenario: very small amounts of low-integrity RNA from FFPE tissue, including single microdissected cells.
Comparison of Smart-3SEQ with selected RNA-seq methods
Cost per library includes all reagents (kits, SPRI beads, enzymes) but not consumables (tubes, pipet tips) and is rounded to the nearest 5 USD.”