Figure 1

Triple match sequencing (TMS): principle and workflow. (A) A complex SAGE-tag based deep sequencing library is generated from total RNA and sequenced on a high-throughput platform (e.g. ABI-SOLiD, ABI-Proton, Illumina) for transcript quantification based on read frequency. A second normalized SAGE library from the same RNA and sequenced on a long read, low-throughput platform (e.g. Roche 454). mRNA reads from these libraries share the distal 3^′-sequence. The long reads are compared to related transcript databases (i.e. human, mouse, rat, dog; RefSeq) as detailed in Methods. (B) Deep sequencing libraries from total RNA were constructed for either SOLiD or 454 sequencing platforms. To ensure representation of low-abundance transcripts it is essential to normalize the 454 library by competitive hybridization. The long 454 reads are assembled into contigs and mapped to human, rat, mouse and dog RefSeq databases yielding about 10’800 annotated genes (hamster liver transcriptome). 93% of 82 million short SAGE tags mapped to the hamster transcriptome and the remaining 7% to human RefSeq entries allowing quantification over 4 orders of magnitude. All 454 and SOLiD sequencing data generated here are available on request.