Development of a high-resolution surveillance protocol using eDNA for detection of brown marmorated stink bugs

Final report for GNE15-112

Project Type: Graduate Student
Funds awarded in 2015: $14,999.00
Projected End Date: 12/31/2017
Grant Recipient: Rutgers University
Region: Northeast
State: New Jersey
Graduate Student:
Faculty Advisor:
Dr. Julie Lockwood
Rutgers University
Faculty Advisor:
Dr. Dina Fonseca
Rutgers University
Expand All

Project Information


It is well established that early detection is the single-most effective strategy for control of invasive species. The brown marmorated stinkbug (BMSB, Halyomorpha halys), which has become a devastating pest to many growers within the mid-Atlantic United States, was first detected in the US in 1996 and is currently still expanding across the continent. Current efforts at monitoring for the presence of BMSB on farms rely on capturing individuals using black light and pheromone traps, which may be producing false negatives while populations are at low abundance; thus delaying management actions that could otherwise effectively control populations on agricultural fields. However, emerging surveillance techniques utilizing a genetic resource known as environmental DNA (eDNA) have a proven track record in aquatic systems, which determines presence of target species at abundances far below what direct monitoring can accomplish. I have adapted these aquatic eDNA techniques for terrestrial uses to develop a spatially fine resolution detection protocol for agriculture pests, specifically BMSB. More specifically, I will be testing individual crop surfaces, topsoil beneath crops, and crop wash water to survey for residual DNA deposition from BMSB to determine detection. This will primarily be accomplished by utilizing a genetic assay I, in collaboration with others, developed and tested specifically for BMSB. Results utilizing this assay thus far indicate that BMSB indeed does leave genetic material on crop surfaces that are detectable in both field and laboratory trials. In addition, conducting surveillance in this way greatly outperformed traditional monitoring practices, indicating the ability to detect small populations more effectively than blacklight and pheromone traps. Ultimately the use of eDNA will allow rapid-onset control to be implemented, which will allow small produce farms to maintain their profitability by keeping costs low, due to reduced pesticide use, while experiencing fewer crop losses.


The brown marmorated stink bug (BMSB - Halyomorpha halys) is native to northeast Asia, and was first found in the United States in Allentown, Pennsylvania in 1996. It has since been found in at least 40 US states, Canada, and several European countries, and has caused significant damage to agricultural crops and ornamental plants resulting in millions of USD in economic losses. The range of crops BMSB attacks is extensive, making it a threat to farmers around the world. Due to the severity of the damage it causes, farmers have controlled populations by increasing their frequency and intensity of broad-spectrum insecticide applications. Such applications are known to be disruptive to natural ecosystems, and undermine integrated pest management efforts. My aim was to develop a novel way to conduct surveillance of nascent populations of BMSB using environmental DNA (eDNA) within an agricultural setting, and test whether such a method would be more effective than current practices.

Project Objectives:

1.     Determine if BMSB eDNA can be detected on small produce farms.

- The goal for this part of the project is to determine the most effective sampling strategies that will allow us to detect BMSB eDNA on farms.


a.     Can BMSB eDNA be detected on crop surfaces, or within suspected feeding sites on produce?

- Under this objective our goal was to utilize forensic grade swabs (i.e. sterile and DNA free) to collect DNA that may have been deposited by BMSB while on the surface of crops.

b.     Can BMSB eDNA be found in the top soil directly beneath crops?

- In this objective our goal was to gather a small amount of topsoil from directly underneath crops. The assumption here was that during rain, irrigation, or just gravity that genetic material would fall off the plant and be left behind in the soil.


c.     Can BMSB eDNA be found within the water of crop wash stations?

- In this objective the success of eDNA surveillance in aquatic systems was adopted and adjusted for use in a terrestrial setting in hopes of equivalent success.


2.     Use eDNA to detect BMSB when they are so rare that direct monitoring fails to detect their presence.

- In this section I will be comparing the effectiveness of the eDNA surveillance being developed here, and traditional monitoring tools currently being employed against BMSB (e.g. blacklight traps and pheromone traps).


3.     Document and disseminate an eDNA surveillance protocol for detection of BMSB.


Click linked name(s) to expand/collapse or show everyone's info
  • Andre Cantelmo
  • Dr. Anne Neilsen (Researcher)
  • Dr. Dina Fonseca
  • Dr. Julie Lockwood


Materials and methods:

Target eDNA collection

I used a genetic tool I previously designed for BMSB that is very sensitive to trace amounts of degraded DNA and exclusively targets BMSB (see Valentin et al., 2016 - Pest Management Science - for more information). The BMSB is a sap-feeder that remains on the host plant for extended periods of time, potentially leaving a detectable level of DNA as they feed, defecate, or molt. The crops they feed on are often harvested by farmers and processed before sale. I posited that BMSB DNA could be gathered from the surface of these crops and used as part of a viable surveillance technique.

To test whether BMSB DNA could be collected from water I placed individual BMSB excreta and exuvia in replicates of six and two, respectively, in a liter of deionized water, with two water-only samples acting as negative controls. Following one methodology in the eDNA literature, I used a peristaltic pump and filter membrane combination to remove the DNA from the water. Once DNA collection was complete, I handled filter membranes with flame sterilized tweezers, and extracted DNA from the filter using an affordable and readily available HotSHOT extraction. I tested for the presence of BMSB DNA with the aforementioned BMSB TaqMan assay (for details regarding assessments of specificity, sensitivity, and protocol please refer to Valentin et al., 2016 - Pest Management Science) in replicates of two within a laminar flow hood with a UV light for surface sterilization prior to qPCR setup to ensure a clean working environment.

To test whether BMSB DNA could be recovered from soil, I once again placed individual BMSB excreta within four 10g samples of loamy soil, with one sample not provided a BMSB sample acting as a negative control. I then extracted them using a commercially available soil extraction kit for large soil volume, in addition to an extraction negative control. All extractions were then tested for BMSB DNA using the same qPCR method described above.


BMSB DNA deposition rate

To document how much time an individual BMSB must be present and feeding on a fruit before a detectable level of eDNA was deposited, I conducted a time series experiment placing a single BMSB adult (from a colony maintained at Rutgers) in a small cage containing a single tomato. I allowed individual BMSB to feed on single tomatoes for a period of two, four, six, or eight hours with four replicates of each treatment (in all 16 BMSB and 16 tomatoes were used). While wearing nitrile gloves, I rinsed each tomato in a bucket containing a liter of deionized water (changing gloves between tomatoes), pumped the water to collect the eDNA, then processed and tested the filters as described above. In addition, as controls, I rinsed and filtered water from tomatoes kept in cages without BMSB (two replicates) and from two tomatoes that were not placed in cages. Filter extraction and qPCR processing were identical to the previous experiment.

            I repeated this experiment, except where tomatoes were rinsed they were instead swabbed with sterile self-contained forensic cotton swabs. The swabs were then processed to extract the DNA from the surface identically to how filters were, and then tested via qPCR again in an identical manner.


Development and testing of field protocol

To examine the efficiency of this protocol in locations varying in levels of BMSB infestation I sampled crops from two farms. The first was in New Jersey (NJ) where BMSB is prevalent, and the second in New Hampshwere (NH) where BMSB have not been confirmed as present in agricultural fields but that sits near the edge of the species’ current known range. At both farms, I performed eDNA based surveillance in conjunction with a blacklight trap and four Dead-Inn 4-ft black pyramid traps with Trécé PHEROCON® BMSB (low dose) pheromone lures so that I could directly compare effectiveness at detecting BMSB. I trapped for BMSB and filtered one to two liters of rinse water at both sites in July and August when they are naturally most abundant.


Field-testing at a high BMSB abundance site

In New Jersey I worked in a peach orchard in the Rutgers Agricultural Research Extension Center (RAREC) in Bridgeton, NJ known to harbor large populations of BMSB. While wearing nitrile gloves, I collected five to seven peaches from four different peach trees and washed them in buckets with one liter of deionized water, while still in the field. All peaches from each tree were washed in the same bucket, and each tree had a pyramid trap and pheromone lure directly next to it that had been placed at the start of the season (with lures regularly replaced). Since each tree had its own trap and was considered a separate location within the site, gloves were changed between trees to prevent cross contamination, and buckets representing each tree were kept isolated from each other to assess positive or negative detections by location within each site. The water in each bucket was processed using the pump and filter combination as in the laboratory experiments. Once filtration was completed, I removed the filter membranes from their housing and placed them in 1.5mL microcentrifuge tubes containing molecular grade 100% ethanol for storage and transport to the lab. Filters were handled using flame-sterilized tweezers and processed as previously described immediately upon return to the lab.


Field-testing at an unknown BMSB abundance site

I further tested the performance of the eDNA field surveillance protocol against conventional monitoring methods (i.e. blacklight traps and pheromone traps) at Heron Pond Farm (NH), a diversified vegetable farm that sits near the expanding front of the BMSB geographical range but was not known to be infested. I visited the New Hampshire farm twice, during the first and third weeks of August. I set one blacklight trap powered from a 120v wall outlet, and four black pyramid traps with lures (identical to those used in NJ) spread throughout four fields containing anywhere from one to three different crop varieties each (cucurbits, chard, kale, arugula, tomatoes, and peppers). All traps were run continuously throughout the sampling period. Blacklight traps were inspected each morning, and pheromone traps inspected both in the morning and at various points throughout the harvesting period each day during each week. To ensure the containers being used were not contaminated with BMSB DNA prior to contact with crops, after each container was filled with the farm’s local water supply (river and well water) and readied for use, I filtered one liter of water and tested the filter paper for the presence of BMSB DNA. This effort ensured that any positive identification of BMSB from using these containers was not due to contamination of BMSB DNA from anywhere else except the crops being washed that day. For each water container, after crops were harvested and thoroughly washed, approximately one to two liters of water, depending on amount of suspended materials and subsequent filter saturation, were pumped through the filtration system and processed for eDNA collection. This resulted in seven to thirteen filter samples per day, from nine different crops over eight days (two four-day sampling periods). While some crops harvested and washed were directly next to the pheromone trap in their respective field, most were over fifty meters away. Samples were processed in the field and lab in the same way as during the experiments in New Jersey.

Research results and discussion:


I found that all water and soil samples spiked with BMSB were qPCR positive, and all negative controls were negative. The time series experiments conducted in water resulted in positive detections in the rinse water across all time ranges (i.e. two, four six, and eight hours) indicating that detectable levels of BMSB DNA were deposited after only a few hours, at least under cage conditions. The time series experiments conducted with the forensic swabs were not consistent (clearly positive samples would not provide positive results), and no conclusion as to why was no determined.

            Rinse water of peaches from the four trees on the New Jersey farm tested positive for BMSB DNA during both visits (Final-Report-Table-1a). Pheromone traps located next to each of the trees were also positive, and on a few occasions we observed BMSB nymphs crawling on peaches just before the fruit was collected for processing. All negative controls were negative.

            I found that the eDNA strategy was both effective in the field and more sensitive to smaller populations than the blacklight and the pheromone traps. At the New Hampshire farm, I found evidence of BMSB eDNA on all eight days sampled (Final-Report-Table-1b). Tests of the wash containers prior to washing harvested crops yielded no positive detections, indicating no pre-contamination. The blacklight trap collected a number of different insect species, but not BMSB. The pheromone traps caught a few native stink bug species throughout the sampling period (e.g. green stink bug – Chinavia hilaris), but only one BMSB, a nymph collected on the last day of sampling (Final-Report-Table-1b). Physical detection in the New Hampshire farm provided a visual confirmation of the presence of BMSB. I note that this nymph was found near the end of August, after BMSB populations had the opportunity to grow throughout the season.


Research dissemination

Results and efforts from this project have been submitted, and accepted for publication to Frontiers of Ecology and the Environment under the title "Early detection of terrestrial invasive insect infestations by using eDNA from crop surfaces". In addition, upon release of the paper from Frontiers, all information will be submitted to in order to disseminate the research materials to extension specialists across the region.



I report here on the development and testing of a novel surveillance tool for a terrestrial insect invader using eDNA, and my results provide strong evidence that such an approach can be used successfully within an agricultural setting. This approach provides much greater sensitivity to the presence or absence of brown marmorated stinkbug (BMSB) than the blacklight and pheromone-baited traps evaluated here. Although these traps were originally designed to monitor population abundance within established or spreading BMSB populations and not for surveillance (Nielsen et al., 2013;Short et al., 2017), they are currently the best option available for either purpose.

The key to this eDNA approach is recognizing that individual BMSB naturally gather on fruit and vegetable crops and regularly deposit their DNA as they feed. At the New Hampshire site, a fully operational vegetable farm, I showed that this eDNA surveillance method could be used to detect a nascent BMSB infestation. While the water collection method worked very well, there were inconsistencies with the swabbing method that have yet to be worked out. Additionally, I came to a realization that in order to get an appropriately representative sample of the presence of BMSB DNA on crops, I would need to sample a wide range of individual crops. This would have subjected me to a sampling issue when attempting to detect BMSB DNA, which was something I was attempting to avoid. Thus, the concept of using swabs was abandoned. Additionally, while soil was successful in detecting BMSB DNA, the cost of soil extraction kits made it a cost-ineffective strategy to implement at this time.

Research conclusions:

Our approach has the potential to revolutionize agricultural pest surveillance, although there remains much needed research regarding the ‘ecology’ of eDNA on working farms, the cost effectiveness of eDNA for surveillance, as well as a better understanding of when an eDNA detection heralds an infestation since not all introductions result in establishment. Nevertheless, the growing number of exotic insects that are known to be harmful to agricultural crops makes such research investments worthwhile. My recommendation would be to apply such surveillance efforts within agricultural crop systems widely, in an attempt to mitigate the spread of current exotic pests, as well as rapidly detect new exotic insect pest populations that have entered the United States.


New Hampshire farm update

Upon receiving the results from the low abundance site in New Hampshire, the farmers were immediately notified of the small BMSB population growing in their fields. As a result they have since begun discussions with local extension specialists in order to proactively manage BMSB populations before they become problematic. This proactive approach will allow them to prevent populations from booming and causing significant losses to their crops, to which are sold directly to the local community, resulting in minimal economic losses. Additionally, their proactive approach will allow them to utilize fewer pesticides than they would have had they reactively attempted to manage a larger growing population on their fields. Furthermore, due to the eDNA technique’s success in determining which crop stands contained BMSB DNA signatures, they could focus their efforts on the infected crop stands as well as those directly adjacent, rather than attempting to spray all their fields simultaneously.

Providing information to farmers regarding pest presence and precise locations of detection would not only relieve farmers from the financial burden of spraying all their fields but also result in reduced chemical input into the environment as well. The utility of eDNA based surveillance can prove to be a phenomenal addition to a farmer’s sustainability toolkit, because it would have minimal impacts on current farm practices, reduce monetary and environmental costs of pesticide use, and allow proactive decision making for managing dangerous crop pests on their plots.

Participation Summary
2 Farmers participating in research

Education & Outreach Activities and Participation Summary

8 On-farm demonstrations
1 Published press articles, newsletters
3 Webinars / talks / presentations

Participation Summary:

2 Number of agricultural educator or service providers reached through education and outreach activities
Education/outreach description:

During the course of this project, I have interacted with the farmers that agreed to participate in the project a great deal. While conducting my research they were inquisitive about the process, and asked numerous questions. As a result, I conducted several demonstrations about the process, how sample collections are done, and what the results mean for both the trapping method and the eDNA surveillance effort. The methods and results of this project have been submitted, and accepted, to Frontiers of Ecology and the Environment. A second publication is currently in preparation to be submitted to Molecular Ecology Resources. In addition, I have presented the methods and results of this project in two national meetings (Ecological Society of America annual meeting - Portland, Oregon; Entomological Society of America annual meeting - Denver, Colorado) and a regional meeting (Entomological Society of America Eastern Branch meeting - Newport, Rhode Island).

As of this writing, I have been requested to present my research at a working group meeting centered around brown marmorated stink bug (BMSB), management efforts, and its pest status nationally. My hope is to inform others of conducting surveillance efforts for BMSB by utilizing the eDNA methods I have established through the course of this project.

Project Outcomes

2 Farmers reporting change in knowledge, attitudes, skills and/or awareness
3 Grants applied for that built upon this project
1 Grant received that built upon this project
$95,000.00 Dollar amount of grant received that built upon this project
5 New working collaborations
Project outcomes:

My project has had large impacts on agricultural sustainability in light of exotic insect pests. Prior to my research, the idea of rapid detection of exotic species populations was limited to what could be seen, which unfortunately also meant extirpation/eradication efforts were unlikely to be successful and long term management was necessary. Now that rapid detection can be a reality for nascent exotic insect pest populations, efforts can be more efficient and cost effective, as well as fewer impacts against the farmers. Ultimately, by substantially reducing, or even eliminating, these threats agricultural systems can be more sustainable through the use of less pesticides, seeing less damage to goods, and suffering fewer economic losses.

Knowledge Gained:

While working on this project, my advisors and I have gained a new appreciation for what it means to have sustainable agriculture in the face of invasive insect pests. We knew that they had significant impacts on crop farms, but were not so prepared for the reality of how sensitive the system was to their presence. As a result, the research conducted while on this grant sought to achieve our aims, while being as minimally impactful to farmers as possible in both an operational and economic way.

This project has also led to applications with other invasive insect pests that threaten agricultural systems. Currently, these new applications have brought about a spree of new research questions ranging from new applications to the implications of its use and results. It has also changed my career trajectory substantially, where my goals are now aimed toward academia. My transition to academia has pushed me to make new collaborations, which further enforces my research direction and provides new questions and applications.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.