Start Monday May 2nd at 4pm (CET, Berlin/Paris)

(3 pm London/Lisbon, 10am New-York, 7am San Francisco, 5pm Tel Aviv)

Main Speaker: Isana Veksler-Lublinsky, Ben-Gurion University of the Negev, Israel (Link)

Title: miRNA target prediction through a machine learning lens

Abstract:

MicroRNAs (miRNAs) are small RNA molecules that hybridize to complementary sequences on target mRNAs and repress their translation to proteins or mediate their degradation. Identifying miRNA target sites on mRNAs is a fundamental step in understanding miRNA function. Novel experimental methods, which can produce high-throughput, unambiguous interacting miRNA–target datasets, have pushed the field forward in recent years. However, due to technical challenges involved in the application of the experimental methods, there is a constantly increasing interest in using computational approaches for miRNA target prediction, especially those that are based on machine learning (ML) models. In my talk, I will describe the challenges involved in the application of ML models to miRNA target prediction. In addition, I will show how we have been using classic and deep learning approaches to investigate whether miRNA-target interaction rules are transferable between species.

Short session speaker (8 minutes long): Shani Cohen (PhD student)

Title: Machine learning for predicting bacterial small RNA-target interactions

Abstract:

Bacterial small RNAs (sRNAs) are relatively short non-coding RNA molecules (~50-500 nt) that play a significant role in the regulation of various bacterial functions, such as virulence, environmental sensing, metabolism, and gene expression. Bacterial sRNAs have a wide variety of regulatory mechanisms, including base-paring with target mRNAs. The two major classes of base-pairing sRNA are commonly called cis-encoded and trans-encoded. The cis-encoded sRNAs are transcribed from the strand complementary to the mRNA they regulate, whereas the trans-encoded share only a partial sequence complementarity with their targets and thus may regulate multiple genes. Similar to microRNAs in eukaryotic, the trans-encoded sRNAs modulate the translation, processing, and/or stability of their target mRNAs by short interactions.
Although hundreds of sRNAs have already been identified, the characterization of their regulatory mechanism in different bacterial species is still limited and much dependent on the discovery of their bona-fide mRNA targets. We collected and processed large datasets of E.coli sRNA-mRNA chimeric interactions from recently published high-throughput experiments. We then extracted a variety of features from each interaction and built advanced machine-learning models for predicting sRNA targets. In this talk, I will share the design of our study, the challenges, and preliminary results.