English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Given the increasing availability of RNA-seq data and its complex andheterogeneous nature, there has been growing interest in applying AI/machinelearning methodologies to work with such data modalities. However, becauseomics data is characterized by high dimensionality and low sample size (HDLSS),current attempts at integrating AI in this domain require significant humanguidance and expertise to mitigate overfitting. In this work we look at howtransfer learning can be improved to learn from small RNA-seq sample sizeswithout significant human interference. The strategy is to gain general priorknowledge about a particular domain of data (e.g. RNA-seq data) by pre-trainingon a general task with a large aggregate of data, then fine-tuning to variousspecific, downstream target tasks in the same domain. Because previous attemptshave shown traditional transfer learning failing on HLDSS, we propose toimprove performance by using Guided Transfer Learning (GTL). Collaborating withRobots Go Mental, the AI we deploy here not only learns good initial parametersduring pre-training, but also learns inductive biases that affect how the AIlearns downstream tasks. In this approach, we first pre-trained on recount3data, a collection of over 400,000 mouse RNA-seq samples sourced from thousandsof individual studies. With such a large collection, patterns of expressionbetween the ~30,000 genes in mammalian systems were pre-determined. Suchpatterns were sufficient for the pre-trained AI agent to efficiently learn newdownstream tasks involving RNA-seq datasets with very low sample sizes andperformed notably better on few-shot learning tasks compared to the same modelwithout pre-training.

Using Guided Transfer Learning to Predispose AI Agent to Learn Efficiently from Small RNA-sequencing Datasets