What’s a more sustainable way to combat cacao pests?

A team of De La Salle University researchers led by Dr. Jose Isagani Janairo of the Department of Biology is exploring a machine learning-driven solution to counter a certain type of insect pest that threatens the production of Philippine cacao industry.

The Philippines has all the natural gifts for cacao production: good climate, favorable soil, and strategic location. For all its potential to be a major player in the multi-billion dollar cocoa-chocolate industry, however, the country is still pulling itself up from the great impact of pest infestation that almost wiped out the local cacao industry back in the ‘90s.

Data from the Department of Science and Technology (DOST) show a continuing decline in the volume of cacao production in the country in the past three decades, except for a short uptick from 2017 to 2019. Local farmers would record their lowest production yet in 2023 with only 1,092.27 metric tons of cacao, which was even lower than those of the pandemic years.

To help boost the sector, a team of researchers from De La Salle University (DLSU) embarked on a DOST-funded research to combat the cacao mirid bug (Helopeltis bakeri), one of the major insect pests in the country and in other parts of Southeast Asia.

Dr. Jose Isagani Janairo, full professor at the DLSU Department of Biology and the country’s project lead, shares that their team seeks to integrate machine learning in the discovery of semiochemicals of the cacao mirid bug. “Semiochemicals are chemical signals used by plants and insects to communicate and navigate their surroundings,” he says.

“The exciting part of our study is we’re eavesdropping on plants and insects. By trying to understand the language of chemistry used by these plants and insects, we can send out fake signals that we can use to fool the pest,” he adds.

From synthetic pesticides to a more sustainable approach

The cacao mirid bug is an insect with a very long proboscis, which is used to puncture the cacao pods. “Once the pods have been punctured, they serve as a gateway for other pathogens to enter, and this leads to accelerated pod rotting,” Janairo says. “We cannot use them for anything else. And so this has been our motivation to develop sustainable pest management approaches to the cacao mirid bug.”

Since 2016, Janairo and his research group at the Institute of Biological Control at the DLSU Laguna Campus have been involved in the development of sustainable approaches to cacao pest management. Exploring the field of semiochemistry, they have focused on one of its applications, which is semiochemical-baited traps.

The research involves the identification of a possible compound to which the cacao mirid bug is attracted. A compound acts like a signal that tells the bug there is food to feed on, and it is exclusive to certain bugs or insects. Once the compound is identified, it can then be placed in a trap.

Janairo points out that since chemical signals are very specific, no other insects, animals, or plants will respond to the trap. The project offers a sustainable alternative to common pest management methods, such as using plastic or synthetic pesticides that harm not only other species but the environment as well.

For their predictive model, the research team is building a dataset of candidate compounds from which they will identify the specific compound that will elicit a physiological response to the pest. The team will study, for instance, if the mirid bug responds to a compound that is more hydrophobic.

Because of the tedious work of semiochemical discovery, which involves a lot of trial and error and demands a lot of resources, the machine learning model will come quite handy, Janairo says.

A boost to the farming industry

The research team is composed of experts from various disciplines, among them entomologists, chemists, and agriculturists, all actively collaborating to develop sustainable pest management approaches, not only for cacao but for other crops as well.

For their latest project, they are building on years of previous efforts that had been undertaken with support from the DOST Philippine Council for Agriculture, Aquatic, and Natural Resources Research and Development. Among their major work was the discovery of a compound called beta-caryophyllene as a behaviorally active semiochemical.

“The challenge for us right now is to develop something more potent,” Janairo says. With the integration

of machine learning, the team could now screen thousands of compounds in a much shorter period of time. Beyond the identification of potent compounds, they express hope for the fulfillment of their ultimate goal: “We want to use science to benefit not only the farmers, but the Filipino people in general.”