A technique for gene expression analysis
that builds on the SAGE
approach but avoids conventional sequencing. A practical realisation was recently concocted by Sydney Brenner
The first step is to use our cDNA library to construct a library of microbeads, each bearing either up to 100,000 fluorophore-tagged copies of a single cDNA or no cDNA at all. Suffice to say that this step involves 3-4x10^4 cDNAs, 1.67x10^7 random 32-mer 'address tag' and 'capture tag' sequences, combinatorial synthesis, PCR and microbeads.
After selection for the 1% of microbeads with bound cDNA via a FACS, the library is immobilised in a two-dimensional flow cell, where each bead’s position is recorded with a microscope and CCD (using the fluorophore). Now, we can read 16-20bp characteristic sequences ('signatures') from each cDNA - in parallel! - by rendering 4bp at a time single-stranded and checking each position in turn to see if it is A, T, G or C. Suffice to say that this step involves Dpn, 'initiating adaptors' with BbvI (a type IIs restriction enzyme) sites, 4bp overhangs, 4^4=1024 'adapter oligonucleotides', and 4^2=16 sets of 4^3=64 'decoder oligonucleotides’ with attached phycoerythrin fluorophores.
So...each microbead generates 1+5x(4x4)=81 signals, revealing first its location for subsequent tracking and then a 20bp characteristic snippet of its cDNA's sequence. Corresponding mRNAs are identified via database searches for these signatures. Testing using 269,093 signatures has shown that the process is almost identical in accuracy to SAGE, and further testing using >1,619,000 signatures revealed matching expression levels to those obtained by convential sequencing (with a few distinct exceptions). Current sequencing speed is the same as for the conventional gel technique but the signatures-per-day rate is >10x greater.
**Brenner, S. Gene expression analysis by massively parallel signal sequencing (MPSS) on microbead arrays, Nat. Biotechnol.. 18:630-634 (2000)