Symbiont-Screener: a reference-free filter to automatically separate host sequences and contaminants for long reads or co-barcoded reads by unsupervised clustering
Xu, M.; Guo, L.; Shi, C.; Liu, X.; Chen, J.; Liu, X.; Fan, G.
Show abstract
Decontamination is necessary for eliminating the effect of foreign genomes on the symbiont studies and biomedical discoveries. However, direct extraction of host sequencing reads with no references remains challenging. Here, we present a triobased method to classify the host error-prone long reads or sparse co-barcoded reads prior to assembly, free of any alignments against DNA or protein references. This method first identifies high-confident host reads by haplotype-specific k-mers inherited from parents, and then groups remaining host reads by the unsupervised clustering. Experimental results demonstrated that this approach successfully classified up to 97.38% of the host human long reads with the precision rate of 99.9999%, and 79.95% host co-barcoded reads with the precision rate of 98.36% using an artificially mixed data. Moreover, the tool also exhibited a good performance on the decontamination of the real algae data. The purified reads reconstructed two haplotypes and improved the assembly with larger contig NGA50 value and less misassemblies. Symbiont-Screener can be freely downloaded at https://github.com/BGI-Qingdao/Symbiont-Screener.
Matching journals
The top 7 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- ViWrap: A modular pipeline to identify, bin, classify, and predict viral-host relationships for viruses from metagenomes 92%
- High-throughput generic single-entity sequencing using droplet microfluidics 92%
- A neural network-based framework to understand the Type 2 Diabetes (T2D)-related alteration of the human gut microbiome 91%
Similar papers in this journal
- Genome Enrichment of Rare, Unknown Species from Complicated Microbiome by Nanopore Selective Sequencing 95%
- Ultra-low input single tube linked-read library method enables short-read NGS systems to generate highly accurate and economical long-range sequencing information for de novo genome assembly and haplotype phasing 94%
- Probing the eukaryotic microbes of ruminants with a deep-learning classifier and comprehensive protein databases 93%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.