3 research outputs found
Recommended from our members
Microphone Conversion: Mitigating Device Variability in Sound Event Classification
No full textMicrophone Conversion: Mitigating Device Variability in Sound Event Classification
No full textIn this study, we introduce a new augmentation technique to enhance the
resilience of sound event classification (SEC) systems against device
variability through the use of CycleGAN. We also present a unique dataset to
evaluate this method. As SEC systems become increasingly common, it is crucial
that they work well with audio from diverse recording devices. Our method
addresses limited device diversity in training data by enabling unpaired
training to transform input spectrograms as if they are recorded on a different
device. Our experiments show that our approach outperforms existing methods in
generalization by 5.2% - 11.5% in weighted f1 score. Additionally, it surpasses
the current methods in adaptability across diverse recording devices by
achieving a 6.5% - 12.8% improvement in weighted f1 score.Comment: Accepted to ICASSP 202
FlyPhoneDB2: A Computational Framework for Analyzing Cell-Cell Communication in Drosophila scRNA-Seq Data Integrating AlphaFold-Multimer Predictions
No full textCell-cell communication (CCC) plays a critical role in the physiological regulation of organisms and has been implicated in numerous diseases. Previously, we introduced FlyPhoneDB, a tool designed to explore CCC in Drosophila single-cell RNA-sequencing datasets. The core algorithm of FlyPhoneDB infers tissue-specific signaling events between cell types by calculating cell-cell interaction scores based on curated ligand-receptor (L-R) expression across major signaling pathways. However, the utility of FlyPhoneDB was limited by the relatively small number of available L-R pairs.
Here, we present FlyPhoneDB2, a major upgrade featuring a significantly expanded knowledgebase that includes a greater number of L-R pairs, incorporating annotations from mammalian species and structural predictions from AlphaFold-Multimer. In addition, the algorithm has been optimized for improved performance and more effective noise filtering. New functionalities have also been introduced, such as the addition of downstream reporter genes to evaluate pathway activity, multi-sample CCC comparison, and enhanced visualizations summarizing communication at a network level.
We demonstrate the utility of FlyPhoneDB2 by analyzing whole-body single-nuclei RNA-seq datasets from flies with gut tumors induced by the Yorkie oncogene. We show that FlyPhoneDB2 not only recapitulates established biological insights into the Drosophila Yorkie tumor model, but also identifies novel potential L-R pairs that may play important roles in tumor-induced cachexia. FlyPhoneDB2 is available at https://www.flyrnai.org/tools/fly_phone_v2/
