Limits...
Towards computational prediction of microRNA function and activity.

Ulitsky I, Laurent LC, Shamir R - Nucleic Acids Res. (2010)

Bottom Line: Our analysis is based on a novel compendium of experimentally verified miRNA-pathway and miRNA-process associations that we constructed, which can be a useful resource by itself.Our method also predicts novel miRNA-regulated pathways, refines the annotation of miRNAs for which only crude functions are known, and assigns differential functions to miRNAs with closely related sequences.Applying our approach to groups of co-expressed genes allows us to identify miRNAs and genomic miRNA clusters with functional importance in specific stages of early human development.

View Article: PubMed Central - PubMed

Affiliation: Blavatnik School of Computer Science, Tel Aviv University, Tel Aviv, Israel. ulitsky@wi.mit.edu

ABSTRACT
While it has been established that microRNAs (miRNAs) play key roles throughout development and are dysregulated in many human pathologies, the specific processes and pathways regulated by individual miRNAs are mostly unknown. Here, we use computational target predictions in order to automatically infer the processes affected by human miRNAs. Our approach improves upon standard statistical tools by addressing specific characteristics of miRNA regulation. Our analysis is based on a novel compendium of experimentally verified miRNA-pathway and miRNA-process associations that we constructed, which can be a useful resource by itself. Our method also predicts novel miRNA-regulated pathways, refines the annotation of miRNAs for which only crude functions are known, and assigns differential functions to miRNAs with closely related sequences. Applying our approach to groups of co-expressed genes allows us to identify miRNAs and genomic miRNA clusters with functional importance in specific stages of early human development. A full list of the predicted mRNA functions is available at http://acgt.cs.tau.ac.il/fame/.

Show MeSH

Related in: MedlinePlus

Comparison of methods for detection of enrichment of miRNA targets. (A) For each miRNA family, all the KEGG pathways were tested for enrichment of miRNA targets and ranked in increasing order of P-value. In case of ties, annotations were ranked in decreasing order of z-score. The chart shows the relative position of the compendium function in each list. (B) Average location of the known KEGG pathway in the ranked lists obtained by using FAME and the HG and LLR tests. Error bars represent one standard error. (C) Average location of the known GO ‘biological process’ annotation in the ranked lists of the three methods. (D) Same as C, but taking into account only annotations that were placed in the top 10% by at least one of the methods.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2926627&req=5

Figure 2: Comparison of methods for detection of enrichment of miRNA targets. (A) For each miRNA family, all the KEGG pathways were tested for enrichment of miRNA targets and ranked in increasing order of P-value. In case of ties, annotations were ranked in decreasing order of z-score. The chart shows the relative position of the compendium function in each list. (B) Average location of the known KEGG pathway in the ranked lists obtained by using FAME and the HG and LLR tests. Error bars represent one standard error. (C) Average location of the known GO ‘biological process’ annotation in the ranked lists of the three methods. (D) Same as C, but taking into account only annotations that were placed in the top 10% by at least one of the methods.

Mentions: We first describe the results on KEGG pathways. Using the compendium, we compared FAME with the HG test, and with the log-likelihood ratio (LLR) scores used by Gaidatzis et al. (10). For each miRNA m associated with a KEGG pathway P in the compendium, we ranked all 140 tested KEGG pathways according to the significance of their enrichment with the targets of m (Figure 2A). The success of each method in predicting a specific function was measured by the rank of P in this list. Eighteen compendium miRNA–pathway pairs met the criterion of at least three genes in P being predicted targets of m, and they were ranked by each of the three methods. In six cases the known pathway corresponded to the top FAME prediction, compared to just four cases when the HG test was applied, and three cases when the LLR test was used (Figure 2A). The average position of the known function across all the 18 pairs was higher for FAME than for the HG and LLR tests (Figure 2B), although the difference was not statistically significant, perhaps due to the small size of the compendium. Performance of the HG test was similar when only the top 25, 50 or 75% of the miRNA–target pairs (as determined by the context score) were used (Supplementary Figure S1A), and it never placed more than four correct pathways as top predictions (results not shown).Figure 2.


Towards computational prediction of microRNA function and activity.

Ulitsky I, Laurent LC, Shamir R - Nucleic Acids Res. (2010)

Comparison of methods for detection of enrichment of miRNA targets. (A) For each miRNA family, all the KEGG pathways were tested for enrichment of miRNA targets and ranked in increasing order of P-value. In case of ties, annotations were ranked in decreasing order of z-score. The chart shows the relative position of the compendium function in each list. (B) Average location of the known KEGG pathway in the ranked lists obtained by using FAME and the HG and LLR tests. Error bars represent one standard error. (C) Average location of the known GO ‘biological process’ annotation in the ranked lists of the three methods. (D) Same as C, but taking into account only annotations that were placed in the top 10% by at least one of the methods.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2926627&req=5

Figure 2: Comparison of methods for detection of enrichment of miRNA targets. (A) For each miRNA family, all the KEGG pathways were tested for enrichment of miRNA targets and ranked in increasing order of P-value. In case of ties, annotations were ranked in decreasing order of z-score. The chart shows the relative position of the compendium function in each list. (B) Average location of the known KEGG pathway in the ranked lists obtained by using FAME and the HG and LLR tests. Error bars represent one standard error. (C) Average location of the known GO ‘biological process’ annotation in the ranked lists of the three methods. (D) Same as C, but taking into account only annotations that were placed in the top 10% by at least one of the methods.
Mentions: We first describe the results on KEGG pathways. Using the compendium, we compared FAME with the HG test, and with the log-likelihood ratio (LLR) scores used by Gaidatzis et al. (10). For each miRNA m associated with a KEGG pathway P in the compendium, we ranked all 140 tested KEGG pathways according to the significance of their enrichment with the targets of m (Figure 2A). The success of each method in predicting a specific function was measured by the rank of P in this list. Eighteen compendium miRNA–pathway pairs met the criterion of at least three genes in P being predicted targets of m, and they were ranked by each of the three methods. In six cases the known pathway corresponded to the top FAME prediction, compared to just four cases when the HG test was applied, and three cases when the LLR test was used (Figure 2A). The average position of the known function across all the 18 pairs was higher for FAME than for the HG and LLR tests (Figure 2B), although the difference was not statistically significant, perhaps due to the small size of the compendium. Performance of the HG test was similar when only the top 25, 50 or 75% of the miRNA–target pairs (as determined by the context score) were used (Supplementary Figure S1A), and it never placed more than four correct pathways as top predictions (results not shown).Figure 2.

Bottom Line: Our analysis is based on a novel compendium of experimentally verified miRNA-pathway and miRNA-process associations that we constructed, which can be a useful resource by itself.Our method also predicts novel miRNA-regulated pathways, refines the annotation of miRNAs for which only crude functions are known, and assigns differential functions to miRNAs with closely related sequences.Applying our approach to groups of co-expressed genes allows us to identify miRNAs and genomic miRNA clusters with functional importance in specific stages of early human development.

View Article: PubMed Central - PubMed

Affiliation: Blavatnik School of Computer Science, Tel Aviv University, Tel Aviv, Israel. ulitsky@wi.mit.edu

ABSTRACT
While it has been established that microRNAs (miRNAs) play key roles throughout development and are dysregulated in many human pathologies, the specific processes and pathways regulated by individual miRNAs are mostly unknown. Here, we use computational target predictions in order to automatically infer the processes affected by human miRNAs. Our approach improves upon standard statistical tools by addressing specific characteristics of miRNA regulation. Our analysis is based on a novel compendium of experimentally verified miRNA-pathway and miRNA-process associations that we constructed, which can be a useful resource by itself. Our method also predicts novel miRNA-regulated pathways, refines the annotation of miRNAs for which only crude functions are known, and assigns differential functions to miRNAs with closely related sequences. Applying our approach to groups of co-expressed genes allows us to identify miRNAs and genomic miRNA clusters with functional importance in specific stages of early human development. A full list of the predicted mRNA functions is available at http://acgt.cs.tau.ac.il/fame/.

Show MeSH
Related in: MedlinePlus