gms | German Medical Science

67. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS), 13. Jahreskongress der Technologie- und Methodenplattform für die vernetzte medizinische Forschung e. V. (TMF)

21.08. - 25.08.2022, online

Combining the results of several PAR-CLIP applications to improve the prediction of microRNA-mRNA interactions using a Bayesian hierarchical model

Meeting Abstract

  • Eva-Maria Hüßler - Institut für Medizinische Informatik, Biometrie und Epidemiologie, Universitätsklinikum Essen, Essen, Germany
  • Martin Schaefer - Deutsches Rheuma-Forschungszentrum Berlin, Berlin, Germany
  • Pablo Landgraf - Universitätsklinikum Köln, Köln, Germany
  • Holger Schwender - Mathematisches Institut, Heinrich Heine Universität Düsseldorf, Düsseldorf, Germany

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie. 67. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e. V. (GMDS), 13. Jahreskongress der Technologie- und Methodenplattform für die vernetzte medizinische Forschung e.V. (TMF). sine loco [digital], 21.-25.08.2022. Düsseldorf: German Medical Science GMS Publishing House; 2022. DocAbstr. 92

doi: 10.3205/22gmds113, urn:nbn:de:0183-22gmds1136

Veröffentlicht: 19. August 2022

© 2022 Hüßler et al.
Dieser Artikel ist ein Open-Access-Artikel und steht unter den Lizenzbedingungen der Creative Commons Attribution 4.0 License (Namensnennung). Lizenz-Angaben siehe



Introduction: MicroRNAs are small non-coding RNAs that play an important role in gene regulation by binding to target mRNA to initiate translational repression and mRNA destabilization. Such targeted mRNA regions can be detected using PAR-CLIP (Photoactivatable-Ribonucleoside-Enhanced Crosslinking and Immunoprecipitation). The transition of T-to-C in the sequenced cDNA induced by this biotechnology helps to distinguish potential binding sites from noise. Only few statistical methods have been developed to detect potential binding sites in PAR-CLIP data.

PAR-CLIP is a method which is applied separately on every sample of an experiment. Bioinformatic or statistical methods are then used to identify binding sites based on the sequenced PAR-CLIP data from a single sample. However, samples are often measured several times so that replicates exist with separate PAR-CLIP applications. Up to now, none of the PAR-CLIP specific methods allows a combined examination of all replicates for a more precise prediction of binding sites.

Methods: We have developed BayMAP [1], [2], a fully Bayesian hierarchical model to detect binding sites. BayMAP is the first method for the analysis of PAR-CLIP data allowing the incorporation of additional information relevant for the biology of microRNA binding sites such as the mRNA region. When using BayMAP the information of as many PAR-CLIP applications as available can be combined by estimating the posterior odds knowing all of these data [2].

Results: In an extensive simulation study and applications to real PAR-CLIP data, BayMAP outperforms other existing methods especially in terms of accuracy and specificity [1], [2]. A simulation study with simulated replicates of a PAR-CLIP experiment reveals, that, when combining results, BayMAP’s sensitivity can be highly improved without decreasing the specificity [2]. An application to real PAR-CLIP applications with replicates indicates that the combined posterior odds are especially useful for positions for which the separate analyses lead to different results. However, the combined posterior odds even changed the prediction for few positions where the separate analyses were unambiguous, due to a larger impact of the data when combing results [2]. Thus, through the combined analysis, it is possible to detect binding sites that would otherwise not have been detected in separate analyses.

Discussion and conclusion: BayMAP is the first PAR-CLIP specific method that allows the combination of several PAR-CLIP applications. Together with BayMAP’s high specificity, this permits to only choose relevant binding sites for further experimental validation in the laboratory. In this talk, we will present the method of the combined posterior odds with its applications to simulated and real PAR-CLIP data.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.


Huessler EM, Schäfer M, Schwender H, Landgraf P. BayMAP: a Bayesian hierarchical model for the analysis of PAR-CLIP data. Bioinformatics. 2019;35(12):1992-2000. DOI: 10.1093/bioinformatics/bty904 Externer Link
Hüßler EM. Detecting binding sites in PAR-CLIP data using a Bayesian hierarchical model [dissertation]. Heinrich Heine University Düsseldorf; 2021.