April 22: European Molecular Biol Lab (EMBL) and Istituto Italiano di Tecnologia RNA Flagship (IIT)

April 22: European Molecular Biol Lab (EMBL) and Istituto Italiano di Tecnologia RNA Flagship (IIT)

🎙 Iuliia Kotova, Giorgio Bini, Gian Gaetano Tartaglia 👥 2K 📅 April 23, 2026 ⏱ 60 min 👁 176 📄 original study 🧭 2026-08-16
Available in: English (current) Français

Keywords

paraspecklesinfluenza A virusRNA-RNA interactionsdeep learningcrosslinking

Summary

The seminar features two talks. First, Iuliia Kotova presents her work on how influenza A virus disrupts paraspeckles to access host RNA-binding proteins. Using in-cell crosslinking and mass spectrometry, she identifies over 100 viral-host protein interactions, many enriched in paraspeckles. She shows that viral proteins NP and NS1 interact with paraspeckle components, and the endonuclease PA-X degrades NEAT1 RNA, leading to paraspeckle disassembly. This disruption releases proteins like SFPQ and HNRNPK that are proviral, enhancing viral replication. Second, Giorgio Bini presents his deep learning approach to predict RNA-RNA interactions. He emphasizes the importance of high-quality experimental datasets, combining UV-crosslinking and protein-based methods to create a training set of 118 high-confidence interactions. His model, likely based on sequence and structure features, aims to improve prediction accuracy for long RNA molecules. The talks highlight the interplay between experimental and computational biology in understanding RNA-mediated processes.

143 words

Critical Evaluation

Value of the Information & Strength of the Argument

The first talk provides valuable insights into the molecular mechanisms of influenza A virus infection, specifically the disruption of paraspeckles. The argumentation is solid, supported by multiple experimental approaches: crosslinking, mass spectrometry, FISH, knockdown/knockout, and overexpression. The model is coherent and well-supported by data. The second talk addresses a critical need in RNA biology: the lack of suitable datasets for training AI models. The speaker emphasizes the careful curation of experimental data, which is essential for reliable predictions. The argumentation is clear, though the presentation is more high-level and less detailed on the model architecture.

Scientific Rigor, Source Quality, Title Accuracy

The seminar is scientifically rigorous, with speakers presenting original research from reputable institutions (EMBL, IIT). The first talk references a preprint, and the second talk mentions a publication in Nature Communications. The title accurately reflects the content, which is a seminar series. The sources cited are appropriate, though the video description only provides a link to the RNA Society’s seminar series page, not directly to the papers. The speakers demonstrate a strong command of their respective fields.

187 words

Title / Content Match

The title accurately reflects the content, which is a seminar series featuring two talks from EMBL and IIT.

Quality & Reliability

8/10

The seminar presents two original research talks with detailed experimental and computational methods, including crosslinking, mass spectrometry, and deep learning. The findings are based on published or preprint work, and the speakers are affiliated with reputable institutions. However, the video is a seminar recording, not peer-reviewed, and some claims are speculative.

Key Moments

Cited Sources

Concurring Sources

Contribution & Novelties

The seminar presents two novel contributions: (1) a comprehensive map of influenza A virus-host protein interactions in the nucleus, revealing paraspeckle disruption as a key strategy to release proviral factors; (2) a deep learning model for RNA-RNA interaction prediction trained on a carefully curated dataset, addressing a critical bottleneck in the field.

Pour aller plus loin :

  • Paraspeckles — Background on paraspeckle structure and function.
  • Influenza A virus — Overview of influenza A virus biology.
  • RNA-RNA interactions — General concept and methods for studying RNA-RNA interactions.
  • Deep learning in biology — Overview of deep learning applications in biological research.

99 words

Radar Profile

The radar profile shows high scores in information quantity, quality, and reliability, with a slightly lower technical level, indicating a seminar that is rich in content and well-supported, but may require some background knowledge to fully appreciate.

Reliability 8/10

💬 No comments were provided for analysis.