Insights from the past: trends in microbial-based weed biological control research

Authors: Fulcher, Michael; Tancos, Matthew

Submitted to: American Phytopathological Society Potomac Division Meeting
Publication Type: Abstract Only
Publication Acceptance Date: 2/24/2021
Publication Date: N/A
Citation: N/A
DOI: https://doi.org/10.1094/PHYTO-111-9-S1.16

Interpretive Summary:

Technical Abstract: Research on the use of plant pathogens as weed biocontrol agents has grown exponentially over the last 20 years, making comprehensive literature reviews challenging. Summarizing and synthesizing this body of research is further complicated by the diverse range of pathosystems under study, including hosts from 23 plant orders and over 180 pathogen species. Automated analyses (e.g. text mining) can augment traditional literature reviews by condensing large document collections and revealing trends obscured by the volume of information available. However, these methods have not been applied widely in the fields of plant pathology or weed management. With a collection of 748 abstracts on the use of plant pathogens for weed biocontrol, we (i) compared the performance of automatic and manual abstract categorization, (ii) described long-term trends in the field, and (iii) identified opportunities for future research based on gaps in the current literature. Automatic and manual annotation of abstracts by “pathogen type” and “land-use” categories produced identical results in 86% of cases. Despite increasing publication rates, the majority of recent research was confined to a single plant order (Asterales) and dealt only with fungal pathogens. More than half of all abstracts introduced novel pathosystems, and relatively few hosts or pathogens were the subject of sustained research. Only 94 of 519 noxious weeds found in the U.S. appeared in the literature, just three host species were the subject of more than 10 published studies, and all but one of those studies focused exclusively on fungal pathogens. The weed biocontrol potential of bacterial and viral plant pathogens has been explored only in passing, and three of the largest orders of invasive plants (Poales, Fabales, Lamiales) have been understudied relative to the number of invasive species they contain. Automated text analysis was an effective way to summarize decades of pathogen-based biocontrol research and could be effectively applied to the review of other sub-disciplines.