The de novo transcriptome sequencing of a weedy plant using GS-FLX 454 technologies is reported. Horseweed (Conyza canadensis L.) was the first broadleaf weed to evolve glyphosate resistance in agriculture, and also is the most widely distributed glyphosate-resistant weed in the United States and the world. However, available sequence data for this species are scant. The transcriptomic sequence should be useful for gene discovery, and to help elucidate the non-target-based glyphosate resistance mechanism and the genomic basis of weediness. Sequencing experiments yielded 411 962 raw reads, an average read length of 233 bp and a total dataset of 95.8 Mb (NCBI accession number SRA010952). After trimming and quality control, 379 152 high-quality sequences were retained and assembled into contigs. The assembly resulted in 31 783 unique transcripts, including 16 102 contigs and 15 681 singletons. The average coverage depth for each contig and each nucleotide position was 22-fold and 12-fold respectively. A total of 16 306 unique sequences were annotated by searching a custom plant protein database. The utility of the transcriptome data was demonstrated by further exploration of ABC transporters, which were previously hypothesized to play a role in non-target glyphosate resistance. Real-time RT-PCR primers were designed from the transcriptome data, which made it possible to assess expression patterns of 17 ABC transporters from resistant and susceptible horseweed accessions from Tennessee, with and without glyphosate treatment. These results show that GS-FLX 454 sequencing is a powerful and cost-effective platform for the development of functional genomic tools for a weed species.