Improving SAGE di-tag processing
© BioMed Central Ltd 2001
Received: 13 February 2001
Published: 22 February 2001
SAGE is a genome-wide method for obtaining gene expression profiles. It generates tags of 10 nucleotides in length, which are assumed to determine the corresponding gene transcript. In practice however, this is not always sufficient for uniquely identifying a gene.
We propose an improved processing of SAGE sequences that allows us to obtain one extra base for reasonably abundant tags. This method includes a statistical test for controlling the relevance of extra base predictions.
The improved SAGE sequence processing we present reduces the uncertainty in SAGE tag to gene mapping and can be applied to any SAGE library.
Serial Analysis of Gene Expression (SAGE) is a method for measuring the relative abundance of gene transcripts in different mRNA samples. It identifies a short mRNA tag from each individual transcript and concatenates them into long DNA molecules, which are then sequenced. By counting these tags one can estimate, for example, the expression of genes in a cell . SAGE popularity is growing fast and many public data are accessible from the Internet .
Di-tag length distribution
Results and Discussion
We use di-tags of sufficient length to compute the frequencies of the four possible extra bases (A, C, G and T) for every tag. Then we use contingency tables and hypothesis testing  to determine relevant extra bases. The null hypothesis we apply is that every possible extra base has the same probability to be sequenced.
Number of 11th base prediction
Number of predictions
For illustration purpose, we identified these tags by extracting SAGE tags of UniGene  clusters (build 108). We only considered tags at the end of the UniGene sequences, i.e. we consider UniGene sequences as 5' oriented. Other identification strategies are possible, see for instance . An example of a tag is CAAGCATCCC, observed 1510 times with 5 extra As (the base A was observed at the 11th position in the di-tag), 1426 extra Cs, 13 extra Gs and 13 extra Ts. We uniquely identified this tag in UniGene as Hs.250444 small inducible cytokine A7 (monocyte chemotactic protein 3). The 11th base found in the UniGene cluster sequence matches with the dominant extra C we mention above. According to the method we propose (see Materials and Methods), the prediction of C as an extra base is relevant at the 99.9% level.
An example of a tag shared by two genes, one of which is apparently not expressed, is provided by GGGCTGGGGT, observed 86 times with 5 extra As and 80 extra Cs. GGGCTGGGGTA is identified in UniGene  as Hs.90436 sperm acrosomal protein (SPAG7) and GGGCTGGGGTC as Hs. 183 698 ribosomal protein L29 (RPL29). According to the extra bases observed, it seems that only RPL29 is expressed (99.9% relevant, SPAG7 is possibly weakly expressed).
The special situation of several expressed genes sharing the same 10-base tag is illustrated by GTGAAACCCC, observed 422 times with 161 extra As, 22 extra Cs, 202 extra Gs, and 12 extra Ts. According to the null hypothesis (equiprobability of every extra base), both A and G are relevant at a higher probability than 95% (99.9% in this case). We can estimate a count of 161/(161+202).422=187 for GTGAAACCCCA and 202/(161+202).422=235 for GTGAAACCCCG. We subsequently found that this tag is shared by many UniGene clusters: 49 clusters with extra A, 7 cluster with extra C and 54 clusters with extra G.
 proposes the assignment of a score to each identification, in order to characterize its reliability. If a tag comes with a predicted extra base, the latter should be checked with the database sequence and the result included in the score computation.
The complex situation of tag GTGAAACCCC above suggests a possible extension of our method. We test the relevance of predicted extra bases by comparing (hypothesis testing) the observed frequencies with the hypothetical situation of equiprobability. Another possible null hypothesis would be (1) to chose a method for identifying tags, as we did with UniGene , and (2) to estimate the relevance of the possible extra bases according to this new null hypothesis. Returning to the example of tag GTGAAACCCC, none of the four possible extra bases significantly departs from the distribution obtained from UniGene. This implies that no extra base can be selected reliably and, consequently, every possible extra base should be considered. We cannot obtain any simplification of the data in that case, contrary to what we found with the equiprobability null hypothesis.
We do not apply the latter extension of the method in practice for two reasons: first, this extension is dependant on the method for identifying tags and, second, considering the difficulty in analyzing SAGE data, we prefer to concentrate on dominantly abundant extra bases for the sake of simplicity.
We presented a method that allows for the prediction of one extra base for sufficiently abundant tags (at least 7 occurrences). The method applies to every SAGE library, without any special preparation. The predictions may be controlled in terms of relevance by using appropriate hypothesis testing techniques. The longer tags permit a better identification of expressed genes.
Materials and Methods
For each di-tag d of length k:
We chose a degree of relevance, typically 95% or 99%. Then, for each different tag t, which has at least one of its extra base counter different from 0, we test whether each possible extra base is relevant (it is possible that more than one extra base is relevant). This is achieved by using contingency tables and hypothesis testing .
Chi-squared statistics allows estimation of the significance of the departure from the null hypothesis. This can be done, for instance, by using the Chi-squared distribution with 1 degree of freedom or Fisher's exact test as soon as D < 5, see .
In our algorithm we only consider di-tags of a length of at least 22 for extra base prediction. We do not use 21-base long di-tags for extra base prediction because (1) the distribution of di-tag lengths (Table 1) shows that there are enough 22-base long di-tags, and (2) this would generate too many wrong 11th base counts, hence making the application of hypothesis testing more difficult.
The authors would like to acknowledge Mark Iberson for reading an early version of this paper. We also thank Massimo de Francesco for his help and his support.
- Velculescu VE, Zhang L, Vogelstein B, Kinzler KW: Serial Analysis of Gene Expression. Science. 1995, 270: 484-487.PubMedView ArticleGoogle Scholar
- Lash AE, Tolstoshev CM, Wagner L, Schuler GD, Strausberg RL, Riggins GJ, Altschul SF: SAGEmap: A Public Gene Expression Resource. Gen. Res. 2000, 10: 1051-1060. 10.1101/gr.10.7.1051.View ArticleGoogle Scholar
- Margulies EH, Innis JW: eSAGE: Managing and Analyzing Data Generated with Serial Analysis of Gene Expression (SAGE). Bioinformatics. 2000, 16(7): 650-651. 10.1093/bioinformatics/16.7.650.View ArticleGoogle Scholar
- Everitt BS: The Analysis of Contingency Tables. London: Chapman and Hall,. 1977Google Scholar
- CGAP (Cancer Genome Anatomy Project). [http://www.ncbi.nlm.nih.gov/CGAP/]
- SAGE library for human normal white matter. [ftp://ncbi.nlm.nih.gov/pub/sage/extr/SAGE_BB542_whitematter/]
- Schuler GD, Boguski MS, Stewart EA, Stein LD, Gyapay G, Rice K, White RE, Rodriguez-Tome P, Aggarwal A, Bajorek E, Bentolila S, Birren BB, Butler A, Castle AB, Chiannilkulchai N, Chu A, Clee C, Cowles S, Day PJ, Dibling T, Drouot N, Dunham I, Duprat S, East C, Hudson TJ, et al: A Gene Map of the Human Genome. Science. 1996, 274(5287): 540-546. 10.1126/science.274.5287.540.View ArticleGoogle Scholar