BodyMap: a collection of 3' ESTs for analysis of human gene expression information.

BodyMap is a collection of site-directed 3' expressed sequence tags (ESTs) (gene signatures, GSs) that contains the transcript compositions of various human tissues and was the first systematic effort to acquire gene expression data. For the construction of BodyMap, cDNA libraries were made, preserving abundance information and histologic resolutions of tissue mRNAs. By sequencing 164,000 randomly selected clones, 88,587 GSs that represent chromosomally coded transcripts have been collected from 51 human organs and tissues. They were clustered into 18,722 independent 3' termini from transcripts, and more than 3000 of these were not found among ESTs assembled in UniGene (Build 75). Assessment of the prevalence of polyadenylation signals and comparison with GenBank cDNAs indicated that there was no significant contamination by internally primed cDNAs or genomic fragments but that there was a relatively high incidence (12%) of alternative polyadenylation sites. We evaluated the sensitivity and resolution of expression information in BodyMap by in silico Northern hybridization and selection of tissue-specific gene probes. BodyMap is a unique resource for estimation of the absolute abundance of transcripts and selection of gene probes for efficient hybridization-based gene expression profiling.

An expression profile of genes in human retina and isolation of a complementary DNA for a novel rod photoreceptor protein.

PURPOSE: To characterize expression patterns of active genes in human retina, and to isolate novel genes that are uniquely expressed in this tissue. METHODS: A 3'-directed complementary DNA (cDNA) library that faithfully represents the composition of messenger RNA (mRNA) was constructed with an mRNA preparation from a cadaveric human retina. A total of 925 3' terminal sequences were collected by sequencing randomly selected clones, of which 789 were regarded as representing chromosomally coded genes (gene signatures [GS]). GS were compared with each other and searched against GenBank. The resulting expression profile, listing gene species and their frequency, represents the composition of mRNA in the retina. By comparing this expression profile with those obtained from 10 other source cells or tissues, genes uniquely active in the retina were discovered, including some not previously described. A full-sized cDNA corresponding to one of these was isolated and sequenced. Its expression was analyzed by multitissue Northern hybridization and in situ hybridization to the retina specimen. It was then mapped on human chromosomes. RESULTS: In the expression profile, 108 genes were detected recurrently, suggesting that they are very active. Fifty-five of them were identified in GenBank, including the most abundant opsin gene and several other genes for phototransduction. Among the remaining novel and active genes, 19 were considered unique to retina on the basis of their representation status in other expression profiles and in dbEST. One of these was identified as a gene that encodes a novel secretory protein expressed in a rod photoreceptor that maps to chromosome 18p11.3. CONCLUSIONS: The expression profile of active genes in the retina represents the composition of mRNA, which reflects the relative activities of genes in this tissue. A comparison of this expression profile with those obtained with other tissues resulted in isolation of a novel cDNA specifically expressed in the rod photoreceptor. It is anticipated that additional novel genes that are uniquely active in the neural retina may be obtained with the same strategy, leading to further clarification of the biologic or physiological characteristics of this tissue.

A gene expression profile of human corneal epithelium and the isolation of human keratin 12 cDNA.

PURPOSE: To describe the quantitative and qualitative aspects of gene expression in human corneal epithelium and to discover novel cornea-specific genes. METHODS: A 3'-directed cDNA library was constructed with messenger RNA prepared from normal human corneal epithelial cells, and inserts in 1069 randomly chosen clones were sequenced. These sequences were compared with each other to determine the frequency of appearance and were searched against GenBank for identification. The resultant expression profile, a list of gene species and their recurrences, reflected the composition of mRNA in the cornea. Recurrently appearing sequences, representing abundant transcripts, were compared with sequences in expression profiles obtained from seven other tissues and from those in dbEST to discover cornea-specific genes. RESULTS: The expression profile of human corneal epithelium showed that the most abundant transcript in this tissue was that for apolipoprotein J. Altogether 62 genes were suggested to be very active, including calcyclin, alpha-enolase, keratin 3, connexin 43, and 12 novel genes. The expression of four of these 12 novel genes seemed to be limited to cornea because they were not found in seven other expression profiles nor in dbEST. Full-length cDNA corresponding to one of these (GS8025), isolated from a separately made cDNA library, contained open reading frame highly homologous to mouse keratin 12, which is known to be cornea specific. CONCLUSIONS: An expression profile of corneal epithelium provides probes to monitor physiological and pathologic conditions of this tissue in terms of gene expression. Furthermore, by comparing this profile with those of other tissues, probes to isolate genes uniquely transcribed in corneal epithelium are determined. These genes are assumed to carry unique functions for this tissue and are candidate genes for inherited diseases that manifest only in cornea. As an example, human cornea-specific keratin was isolated, and partial cDNA sequences for three more cornea-specific genes were presented.

Isolation and chromosomal localization of the human cone cGMP phosphodiesterase gamma cDNA (PDE6H).

By comparing lists of 3'-directed partial cDNA sequences (gene signatures) randomly collected from various tissues, genes unique to each tissue can be identified. A full-size cDNA clone, corresponding to one such tissue-specific gene signature that recurred only in a retina library, was isolated and analyzed. This clone encoded 83 amino acids highly homologous to the bovine blue cone cGMP phosphodiesterase gamma subunit. The retina-specific expression of this gene was confirmed by multiple tissue Northern blotting, and its cone specificity was confirmed by in situ hybridization to a human retina specimen. From these results, we concluded that this clone (HGMW-approved symbol PDE6H) encodes for the gamma subunit of human cone-specific cGMP phosphodiesterase, and it was assigned to chromosome 12p13 by fluorescence in situ hybridization.