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_27
Snippet: We compare the networks to the k-NN classifier proposed by Zhang et al. (2019) and used by them for read-based predictions. We trained the classifier on the "All" dataset as described by the authors, i.e. using non-overlapping, 500bplong contigs generated from the training genomes (retraining on simulated reads is computationally prohibitive). We also tested the performance of using BLAST to search against an indexed database of labeled genomes. .....
Document: We compare the networks to the k-NN classifier proposed by Zhang et al. (2019) and used by them for read-based predictions. We trained the classifier on the "All" dataset as described by the authors, i.e. using non-overlapping, 500bplong contigs generated from the training genomes (retraining on simulated reads is computationally prohibitive). We also tested the performance of using BLAST to search against an indexed database of labeled genomes. We constructed the database from the "All" training set and used discontiguous megablast to achieve high inter-species sensitivity.
Search related documents:
Co phrase search for related documents- network compare and training set: 1, 2
- non overlapping and training set: 1
- read base and simulated read: 1
Co phrase search for related documents, hyperlinks ordered by date