Biology subjects

Zha, Y.

Publications and source records attributed to Zha, Y..

3 recordsLinked to original sources

bioRxiv2021

Enabling technology for microbial source tracking based on transfer learning: From ontology-aware general knowledge to context-aware expert systems

Microbial source tracking quantifies the potential origin of microbial communities, facilitates better understanding of how the taxonomic structure and community functions were formed and maintained. However, previous methods involve a tradeoff between speed and accuracy, and have encountered difficulty in source tracking under many context-dependent settings. Here, we present EXPERT for context-aware microbial source tracking, in which we adopted a Transfer Learning approach to profoundly elevate and expand the applicability of source tracking, enabling biologically informed novel microbial knowledge discovery. We demonstrate that EXPERT can predict microbial sources with performance superior to other methods in efficiency and accuracy. More importantly, we demonstrate EXPERTs context-aware ability on several applications, including tracking the progression of infant gut microbiome development and monitoring the changes of gut microbiome for colorectal cancer patients. Broadly, transfer learning enables accurate and context-aware microbial source tracking and has the potential for novel microbial knowledge discovery.

bioinformatics↗

bioRxiv2020

Ontology-Aware Deep Learning Enables Ultrafast, Accurate and Interpretable Source Tracking among Sub-Million Microbial Community Samples from Hundreds of Niches

The taxonomical structure of microbial community sample is highly habitat-specific, making it possible for source tracking niches where samples are originated. Current methods face challenges when the number of samples and niches are magnitudes more than current in use, under which circumstances they are unable to accurately source track samples in a timely manner, rendering them difficult in knowledge discovery from sub-million heterogeneous samples. Here, we introduce a deep learning method based on Ontology-aware Neural Network approach, ONN4MST (https://github.com/HUST-NingKang-Lab/ONN4MST), which takes into consideration the ontology structure of niches and the relationship of samples from these ontologically-organized niches. ONN4MSTs superiority in accuracy, speed and robustness have been proven, for example with an accuracy of 0.99 and AUC of 0.97 in a microbial source tracking experiment that 125,823 samples and 114 niches were involved. Moreover, ONN4MST has been utilized on several source tracking applications, showing that it could provide highly-interpretable results from samples with previously less-studied niches, detect microbial contaminants, and identify similar samples from ontologically-remote niches, with high fidelity.

bioinformatics↗

bioRxiv2020

Optimal protocols for sequence-based characterization of the human vaginal microbiome

The vaginal microbiome has been connected to a wide range of health outcomes. This has led to a thriving research environment, but also to the use of conflicting methodologies to study its microbial composition. Here we systematically assess best practices for the sequencing-based characterization of the human vaginal microbiome. As far as 16S rRNA gene sequencing is concerned, the V1-V3 region has the best theoretical properties, but limitations of current sequencing technologies mean that the V3-V4 region performs equally well. Both of these approaches present very good agreement with qPCR quantification of key taxa, provided an appropriate bioinformatic pipeline is used. Shotgun metagenomic sequencing presents an interesting alternative to 16S amplification and sequencing, but it is not without its challenges. We have assessed different tools for the removal of host reads and the taxonomic annotation of metagenomic reads, including a new, easy-to-build and - use, reference database of vaginal taxa. This strategy performed as well as the best performing previously published strategies. Despite the many advantages of shotgun sequencing none of the shotgun approaches assessed here had as good agreement with the qPCR data as 16S rRNA gene sequencing. ImportanceThe vaginal microbiome has been connected to a wide range of health outcomes, from susceptibility to sexually transmitted infections to gynecological cancers and pregnancy outcomes. This has led to a thriving research environment, but also to conflicting available methodologies, including many studies that do not report their molecular biological and bioinformatic methods in sufficient detail for them to be considered reproducible. This can lead to conflicting messages and delay progress from descriptive to intervention studies. By systematically assessing best practices for the characterization of the human vaginal microbiome, this study will enable past studies to be assessed more critically and assist future studies in the selection of appropriate methods for their specific research questions.

molecular biology↗