Efficient targeted transcript discovery via array-based normalization of RACE libraries

RACE (Rapid Amplification of cDNA Ends) is a widely used approach for transcript identification. However, the dynamic range in the population of RACE transcript isoforms may be very large, and random clone selection -the typical approach- may be ineffective in sampling the different transcript species present in the population. Here, we describe an effective RACE sampling strategy. The products of the RACE reaction are hybridized onto high-density tiling arrays, and the exons detected are then used to delineate a series of RT-PCR reactions, through which the original RACE mixture is segregated into a number of simpler RT-PCR reactions. These are independently cloned, and randomly selected clones are sequenced. This approach is superior to the direct cloning and sequencing of the RACE products: it specifically targets novel transcripts, and often leads to the overall normalization of their abundances. We indeed show theoretically that this strategy leads to a very efficient sampling of the novel transcript species associated to annotated loci. In a pilot experiment, we used this approach to discover many novel transcripts for a few otherwise well-characterized protein coding genes. Finally we investigate how this strategy can be multiplexed for large-scale transcript discovery by high-density pooling of RACE reactions prior to hybridization. Our results indicate that through the interrogation of a limited number of exons per gene on a limited number of cell types, it is possible to recover a large fraction of the transcript diversity associated to protein coding loci. These loci, however, could be occupying a much larger genomic space than previously expected, implying that efficient multiplexing requires non-trivial pooling optimization. Question-1 (Q1 Samples): To uncover the ideal number and choice of tissues to get the best coverage of transcription (both annotated and newly-identified exons), we performed 5'- and 3' RACEs of 12 genes on polyA+ RNAs from 48 different cell/tissue types, i.e. 33 human adult tissues (Adrenal Gland, Bladder, Brain, Brain Frontal Lobe, Brain Hippocampus, Brain Hypothalamus, Cerebellum, Colon, Epididymus, Heart, Kidney, Liver, Lung, Lymph Node, Mammary Gland, Muscle, Ovary, Pancreas, Pituitary Gland, Placenta, Prostate, Salivary Gland, Skin, Small Intestine, Spinal Cord, Spleen, Stomach, Testis, Thymus, Thyroid, Tongue, Tonsil, Uterus; all BD Clontech), 9 fetal tissues (Fetal Adrenal Gland, Fetal Brain, Fetal Heart, Fetal Kidney, Fetal Liver, Fetal Lung, Fetal Spleen, Fetal Thymus, Whole Fetus; all BD Clontech) and 6 cell lines (GM06990, HeLaS3, HepG2, HL60, K562, SW480) using the BD SMARTTM RACE cDNA amplification kit (BD Clontech Cat. No.634914). Overall: 96 experiments representing two pools of 6 genes done in the 48 samples separately. Question-2 (Q2 Samples): To define how many exons per gene should be interrogated, we performed 5'RACE experiments in ten exons of a set of 44 genes, each mapping to a different ENCODE region15,23 on polyA+ RNAs from 12 tissues (brain, heart, kidney, spleen, liver, colon, small intestine, muscle, lung, stomach, testis, placenta, all BD Clontech). Overall: 120 experiments representing 10 pools each with RACE reactions from a separate exon of the 44 genes done in the 12 samples separately. Question-3 (Q3 Samples): The same set of 12 RNAs and 5'RACE of 96 genes mapping to HSA21 and HSA22 was used to estimate the genomic space occupied by a gene. Overall: 16 experiments representing 16 pools of 6 genes each done on a pool of 12 samples.