Pseudogenes, alternative transcripts, noncoding RNA, and polymorphisms each add extensive complexity to the mammalian transcriptome and confound estimation of the total number of genes. Despite advanced algorithms for gene prediction and several large-scale efforts to obtain cDNA clones for all human open reading frames (ORFs), no single collection is complete. To enhance this effort, we have developed a high-throughput pipeline for reverse transcription PCR (RT-PCR) gene recovery. Most importantly, novel molecular strategies for improving RT-PCR yield of transcripts that have been difficult to isolate by other means and computational strategies for clone sequence validation have been developed and optimized. This systematic gene recovery pipeline allows both rescue of predicted human and rat genes and provides insight into the complexity of the transcriptome through comparisons with existing data sets.