| 
      
        |  | 
          
            |  |  | 
			
				 
                   
			
                
                  |  |  
                  |  |  
                  | 
              
                    
                      |  | BIO-03  Diversity, dynamics and function of host-associated microbiomes in marine organisms |  
                      |  |  
                      | 
                          
                            |  | Symbiont-Screener: a reference-free tool to separate host sequences
from symbionts for error-prone long reads Monday 9th @ 1030-1050, Conference Room 5
 Mengyang Xu*   , BGI-Qingdao, Qingdao, 266555, China
BGI-Shenzhen, Shenzhen, 518083, China
 Lidong Guo, BGI-Qingdao, Qingdao, 266555, China
College of Life Sciences, University of Chinese Academy of Sciences, Beijing 100049, China
 Chengcheng Shi, BGI-Qingdao, Qingdao, 266555, China
 Xiaochuan Liu, BGI-Qingdao, Qingdao, 266555, China
 Yanwei Qi, BGI-Qingdao, Qingdao, 266555, China
 Jianwei Chen, BGI-Qingdao, Qingdao, 266555, China
 Jinglin Han, BGI-Qingdao, Qingdao, 266555, China
 Li Deng, BGI-Qingdao, Qingdao, 266555, China
 Xin Liu, BGI-Qingdao, Qingdao, 266555, China
BGI-Shenzhen, Shenzhen, 518083, China
State Key Laboratory of Agricultural Genomics, BGI-Shenzhen, Shenzhen, 518083, China
 Guangyi Fan, BGI-Qingdao, Qingdao, 266555, China
BGI-Shenzhen, Shenzhen, 518083, China
State Key Laboratory of Agricultural Genomics, BGI-Shenzhen, Shenzhen, 518083, China
 Presenter Email: xumengyang@genomics.cn
 |  
                            | Although decontamination is necessary for eliminating the effect of foreign genomes on symbiosis research and biomedical discoveries, the direct extraction of host error-prone long reads with no references remains challenging. We here present Symbiont-Screener, a reference-free approach to identify host raw long reads according to a trio-based screening model, which exploits strobemer and unsupervised clustering to overcome high error rates. When applied to simulated and real contaminated datasets, it outperforms other de novo decontamination tools, and obtains high precision and recall of decontamination comparable to that of state-of-the-art reference-based classifiers, thus promising a high-quality genome reconstruction of the host. The code of the analysis is available at https://github.com/BGI-Qingdao/Symbiont-Screener. This research was supported by the National Natural Science Foundation of China (Grant No. 32100514). |  |  |  
                  |  |  |  |  |  |