Genome-Based UniGene Build Procedure
The availability of genomic sequence is helpful for identifying sets of transcript sequences that correspond to distinct transcription loci or to annotated genes, which is the goal of UniGene. The procedure used for genome-based clustering of transcript sequences is described here.
Several types of evidence are used to identify a transcription locus' boundaries and to identify which transcripts represent the locus. Although determining gross structure and transcript representation of many genes is not sensitive to the details of transcript mapping, there are cases where the details are important: overlapping genes on opposite strands, or genes located within introns of other genes, for example. To accurately resolve these and similar cases, we identify genes by incorporating evidence in order of confidence, beginning with the strongest data.
Annotation of characterized genes annotated on the genomic sequence is recorded. Annotated genes include those supported by experimentally confirmed RefSeqs as well as transcription loci that are predicted to encode a protein realized in a gene model. These annotated exon boundaries and the association of exons with genes form a skeleton that can be extended by subsequent analysis but cannot be contradicted subsequently.
Transcribed sequences that can be stringently aligned to genomic sequence with a requirement of splice site consensus are used to enumerate additional exon-intron boundaries. Not all sequence alignments satisfy this stringent requirement. Any sequences sharing an exon-intron boundary that can be identified with only one gene are grouped together.
Unspliced sequences, as well as sequences for which the splicing location or orientation is uncertain, are associated with an overlapping exon if one exists, or placed against genome if not. Sequence orientation is used where there is possible ambiguity of gene orientation.
Sequences that do not align to genomic sequence are grouped together, and transcribed sequences within an interval smaller than 3000 nt that have a common clone of origin are grouped together.
Clusters that do not correspond to an annotated gene and are less than 500 bases 3' of another cluster are likely alternative 3' termini, and are merged into the upstream cluster. This merging is not transitive.
Transcript-Based Build Procedure
Clustering is the process of finding subsets of sequences that belong together within a larger set. This is done by converting discrete similarity scores to Boolean links between sequences. That is, two sequences are considered linked if their similarity exceeds a threshold. UniGene clustering proceeds in several stages, with each stage adding less reliable data to the results of the preceding stage. This staged clustering affords greater control than a more egalitarian treatment of all links between sequences.
Screening for contaminants, repeats, and low-complexity sequence is performed. Low-complexity screening is performed using NCBI's Dust program. Mitochondrial and ribosomal sequences are screened for, as are vector contaminants and repetitive elements. After screening, a sequence must contain at least 100 informative bp to be a candidate for entry into UniGene.
Builds are either genome based or transcript based, as described here.Sequence records in distinct Entrez gene records are used to group mRNAs (please see entrez Gene help for details). Links to HomoloGene clusters are computed for all mRNA sequences via translating searches against proteins in Homologene. The set of mRNA sequences is compared with itself. Sequence pairs that are sufficiently similar are linked together to form initial clusters, as long as sequences in each initial cluster do not correspond to multiple Entrez gene records or multiple HomoloGene entries.
Links between ESTs and mRNA are added to these clusters. The set of ESTs is compared with sequences from the set of initial clusters using megaBLAST, and sufficiently similar sequence pairs are added to the clusters. Links that would join the initial mRNA-based clusters are discarded. EST to EST links are also generated and used to extend the initial clusters and to generate clusters composed solely of ESTs.
Clone-based edges are added; these allow non-overlapping 5' and 3' ESTs to be assigned to the same cluster. Because of imperfect clone labeling, a single clone-ID based edge is insufficient to merge two clusters. Clone IDs that link at least two 5' ends from one cluster with at least two 3' ends from another cluster are found, and the two clusters are merged.
Any resulting cluster that does not contain a sequence with a polyadenylation signal or tail is discarded. Clusters that meet these criteria are called anchored clusters, because their 3' ends are presumed to be known.
ESTs that do not belong to an anchored cluster are rechecked at a lower level of stringency than in the preceding passes. An EST that passes this less stringent test is then added to the cluster that contains the sequence that is the best match to the EST; it is a guest member.
Clusters of size 1 (that is, clusters that seem to identify infrequently expressed genes) are compared against the rest of the sequences in UniGene at a lower level of stringency and merged with the cluster containing the most similar sequence.
The resulting clusters are compared with the preceding week's build and renumbered in an attempt to maintain continuity. Because the sequences that make up a cluster may change from week to week and because the cluster identifier may disappear (typically when two clusters merge), using the cluster identifier as a reference is ill advised. Using the GB accession numbers of the sequences that make up the cluster is a safe alternative.