Author: Jakub M Bartoszewicz; Anja Seidel; Bernhard Y Renard
Title: Interpretable detection of novel human viruses from genome sequencing data Document date: 2020_1_30
ID: ac00tai9_65
Snippet: We presented a new approach for predicting a host of a novel virus based on a single DNA read or a read pair, cutting the error rates in half compared to the previous state-of-the-art. For convolutional filters, we jointly visualize nucleotide contributions and information content. Finally, we use GWPA plots to gain insights into the models' behavior and analyze a recently emerged SARS-CoV-2 virus. Data is available at https://doi.org/10.5281/zen.....
Document: We presented a new approach for predicting a host of a novel virus based on a single DNA read or a read pair, cutting the error rates in half compared to the previous state-of-the-art. For convolutional filters, we jointly visualize nucleotide contributions and information content. Finally, we use GWPA plots to gain insights into the models' behavior and analyze a recently emerged SARS-CoV-2 virus. Data is available at https://doi.org/10.5281/zenodo.3630803 and the code is submitted in Supplementary Material. The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.01.29.925354 doi: bioRxiv preprint
Search related documents:
Co phrase search for related documents- art previous state compare and previous state: 1
- convolutional filter and information content: 1, 2, 3
- convolutional filter and information content nucleotide contribution: 1, 2, 3
- convolutional filter and new approach: 1, 2, 3
- convolutional filter and nucleotide contribution: 1, 2, 3
- dna read and error rate: 1
- dna read and single dna read: 1, 2, 3
- error rate and information content: 1
- error rate and model behavior: 1, 2, 3
- error rate and new approach: 1, 2, 3
- error rate and novel virus: 1, 2
- error rate and read pair: 1, 2
- error rate and SARS virus: 1, 2, 3, 4
- error rate and single dna read: 1
Co phrase search for related documents, hyperlinks ordered by date