[Nobel Prize for RNA interference].
Explore the source record for details and available documents.
History, 21st Century↗
Biomedical subjects
Pål Saetrom
Publications and source records attributed to Pål Saetrom.
Reliable prediction of Drosha processing sites improves microRNA gene prediction.
MOTIVATION: Mature microRNAs (miRNAs) are processed from long hairpin transcripts. Even though it is only the first of several steps, the initial Drosha processing defines the mature product and is characteristic for all miRNA genes. Methods that can separate between true and false processing sites are therefore essential to miRNA gene discovery. RESULTS: We present a classifier that predicts 5' Drosha processing sites in hairpins that are candidate miRNAs. The classifier, called Microprocessor SVM, correctly predicts the processing site for 50% of known human 5' miRNAs, and 90% of its predictions are within two nucleotides of the true site. Another classifier that is trained on the output from the Microprocessor SVM outperforms existing methods for prediction of unconserved miRNAs. Reanalysis of characteristics and supporting evidence for a set of newly annotated miRNAs shows that some miRNAs may be misannotated. This suggests that expressed hairpins should not be annotated as miRNAs until they are verified to be Drosha and Dicer substrates. AVAILABILITY: The classifiers are publicly available at https://demo1.interagon.com/miRNA/
Artificial Intelligence↗
Conserved microRNA characteristics in mammals.
Short hairpin RNAs (shRNAs) and short interfering RNAs (siRNAs) probably enter different stages of the microRNA (miRNA) pathway for depletion of mRNA and suppression of protein translation. Primary and secondary structural characteristics that are shared between endogenous primary miRNA transcripts (pri-miRNAs) may contribute toward efficient biogenesis and potent silencing. This study investigates known miRNA transcripts for characteristics that are conserved between miRNAs and that distinguish them from random hairpins with similar lengths. The primary structure is conserved, as demonstrated by a significant presence or absence of certain bases at specific positions in the miRNA precursors and their flanking regions. The secondary structure is also conserved between miRNAs, as internal loops and bulges commonly appear in specific positions in the miRNA stem. The conservation of base-pairing continues past the mature duplex and 13 bases into the primary stem, with no detectable conservation of secondary structure beyond this region. Based on these observations, we have designed a hairpin construct that incorporates the most important characteristics present in endogenous miRNAs. Preliminary experiments suggest that this construct may rescue the efficacy of shRNA triggers that cannot be used with a miR-30-based hairpin, and vice versa.
Animals↗
Predicting non-coding RNA genes in Escherichia coli with boosted genetic programming.
Several methods exist for predicting non-coding RNA (ncRNA) genes in Escherichia coli (E.coli). In addition to about sixty known ncRNA genes excluding tRNAs and rRNAs, various methods have predicted more than thousand ncRNA genes, but only 95 of these candidates were confirmed by more than one study. Here, we introduce a new method that uses automatic discovery of sequence patterns to predict ncRNA genes. The method predicts 135 novel candidates. In addition, the method predicts 152 genes that overlap with predictions in the literature. We test sixteen predictions experimentally, and show that twelve of these are actual ncRNA transcripts. Six of the twelve verified candidates were novel predictions. The relatively high confirmation rate indicates that many of the untested novel predictions are also ncRNAs, and we therefore speculate that E.coli contains more ncRNA genes than previously estimated.
Algorithms↗
Weighted sequence motifs as an improved seeding step in microRNA target prediction algorithms.
We present a new microRNA target prediction algorithm called TargetBoost, and show that the algorithm is stable and identifies more true targets than do existing algorithms. TargetBoost uses machine learning on a set of validated microRNA targets in lower organisms to create weighted sequence motifs that capture the binding characteristics between microRNAs and their targets. Existing algorithms require candidates to have (1) near-perfect complementarity between microRNAs' 5' end and their targets; (2) relatively high thermodynamic duplex stability; (3) multiple target sites in the target's 3' UTR; and (4) evolutionary conservation of the target between species. Most algorithms use one of the two first requirements in a seeding step, and use the three others as filters to improve the method's specificity. The initial seeding step determines an algorithm's sensitivity and also influences its specificity. As all algorithms may add filters to increase the specificity, we propose that methods should be compared before such filtering. We show that TargetBoost's weighted sequence motif approach is favorable to using both the duplex stability and the sequence complementarity steps. (TargetBoost is available as a Web tool from http://www.interagon.com/demo/.).
5' Untranslated Regions↗
Designing effective siRNAs with off-target control.
Successful gene silencing by RNA interference requires a potent and specific depletion of the target mRNA. Target candidates must be chosen so that their corresponding short interfering RNAs are likely to be effective against that target and unlikely to accidentally silence other transcripts due to sequence similarity. We show that both effective and unique targets exist in mouse, fruit fly, and worm, and present a new design tool that enables users to make the trade-off between efficacy and uniqueness. The tool lists all targets with partial sequence similarity to the primary target to highlight candidates for negative controls.
Algorithms↗
A comparison of siRNA efficacy predictors.
Short interfering RNA (siRNA) efficacy prediction algorithms aim to increase the probability of selecting target sites that are applicable for gene silencing by RNA interference. Many algorithms have been published recently, and they base their predictions on such different features as duplex stability, sequence characteristics, mRNA secondary structure, and target site uniqueness. We compare the performance of the algorithms on a collection of publicly available siRNAs. First, we show that our regularized genetic programming algorithm GPboost appears to have a higher and more stable performance than other algorithms on the collected datasets. Second, several algorithms gave close to random classification on unseen data, and only GPboost and three other algorithms have a reasonably high and stable performance on all parts of the dataset. Third, the results indicate that the siRNAs' sequence is sufficient input to siRNA efficacy algorithms, and that other features that have been suggested to be important may be indirectly captured by the sequence.
Algorithms↗
Predicting the efficacy of short oligonucleotides in antisense and RNAi experiments with boosted genetic programming.
MOTIVATION: Both small interfering RNAs (siRNAs) and antisense oligonucleotides can selectively block gene expression. Although the two methods rely on different cellular mechanisms, these methods share the common property that not all oligonucleotides (oligos) are equally effective. That is, if mRNA target sites are picked at random, many of the antisense or siRNA oligos will not be effective. Algorithms that can reliably predict the efficacy of candidate oligos can greatly reduce the cost of knockdown experiments, but previous attempts to predict the efficacy of antisense oligos have had limited success. Machine learning has not previously been used to predict siRNA efficacy. RESULTS: We develop a genetic programming based prediction system that shows promising results on both antisense and siRNA efficacy prediction. We train and evaluate our system on a previously published database of antisense efficacies and our own database of siRNA efficacies collected from the literature. The best models gave an overall correlation between predicted and observed efficacy of 0.46 on both antisense and siRNA data. As a comparison, the best correlations of support vector machine classifiers trained on the same data were 0.40 and 0.30, respectively.
Algorithms↗