PeerJ Award Winners at 18IMCO

The International Meiofauna Conference (IMCO), first held in 1969, is the flagship conference of the International Association of Meiobenthologists (IAM) and is held every three years. The 18th IMCO was held for the first time in an online format on 5-9 December 2022, in conjunction with an online version of Meioscool (a one-week workshop for students and early career researchers wanting to learn about meiofauna taxonomy, ecology and sampling) the preceding week. Both Meioscool and 18IMCO had approximately 100 registered participants from all around the globe. Thanks to the generous support of the International Seabed Authority, students and researchers from developing countries were given free access to the 18IMCO platform as well as an opportunity to present their research findings. A high diversity of topics were covered during 18IMCO, spanning meiofauna from deep sea, coastal environments and even trees, taxa ranging from nematodes, crustaceans, tardigrades, loriciferans, kinorhynchs and flatworms (among others), and techniques such as automated imaging and identification, single specimen genomics and the revolutionary meioflume! Despite the limitations of the online format, many lively discussions were held both within and outside the conference sessions, which will hopefully inspire many of us to continue exploring new and exciting research areas. 

Daniel Leduc. 2022 Conference Convenor 

 

*****

PeerJ Computer Science Conference Collection:

18th International Meiofauna Conference

 

All presenters and attendees of 18IMCO are invited to submit full research articles on any meiofauna-related topic to this Conference Collection. When submitting, please include “Submitted to the 18IMCO Collection – PJLE-COLL-IMCO”  in the Confidential Note to Staff field of the submission form. Manuscripts will undergo peer review at PeerJ Computer Science, and accepted articles will be subject to publication charges

The deadline for submissions is 19th May 2023.

 

*****

 

Luciana Erika Yaginuma PhD candidate. University of São Paulo, Brazil

Can you tell us a bit about yourself and your research interests?

I have a bachelor’s degree in Oceanography (2008) and master’s in Biological Oceanography (2011) both from the Oceanographic Institute of the University of São Paulo in Brazil. Between graduating in the master’s program and my return to academia in 2019, I have worked in environmental monitoring programs in Brazil and had two lovely kids. My research interests are marine meiofauna and nematodes with emphasis on their interaction with the environment.

What first interested you in this field of research?

When I was a graduate student and started working with meiofauna and nematodes, the first thing that interested me was the diversity in what was just sediment at first glance. These tiny creatures, living in the same environment, were so diverse and successful that I ended up entering the world of meiofauna, although it was laborious.

Can you briefly explain the research you presented at 18IMCO?

The research that I presented at the 18IMCO was about using machine learning to predict spatial distribution of meiofauna for environmental purposes. Meiofauna indicators were modelled in relation to environmental variables and interpolated in a high-resolution grid map from Santos Basin (southeast coast of Brazil). Six benthic zones could be recognized based on the meiofauna descriptors and their response to the environmental conditions. Some zones were better recognized by most meiofauna descriptors and others by specific ones, implying that meiofauna indicators should be monitored concomitantly. Results showed that 15 environmental variables were sufficient to retrieve accurate predictions. These results can support the optimizations of future monitoring programs in order to reduce costs and increase our understanding of the system.

How will you continue to build on this research?

The next steps of our research group is to model nematode genera associations and see what else they can tell us. We also want to add temporal variability to our models. The application “iMESc: An Interactive Machine Learning App for Environmental Science” (DOI 10.5281/zenodo.6484391) that we use, developed by two of our group (Danilo Vieira and Gustavo Fonseca), is under constant improvement and it is intended to dynamically integrate machine learning techniques to explore multivariate data sets. The idea is to expand the use of machine learning techniques for environmental science, since more and more data is available these days.

 

Simone Brito Masters student. Federal University of São Paulo, Brazil.

Can you tell us a bit about yourself and your research interests?

My name is Simone Brito, I am a biologist and Master student in Marine Biodiversity and Ecology at the Federal University of São Paulo, Santos, Brazil. I started working and studying meiofauna ecology and nematode taxonomy 10 years ago. During my career, I had the opportunity to work in several projects with diversified habitats, like mangroves, sandy beaches, deep-sea and estuaries. Through those experiences I acquired an integrative vision of the meiofauna and its function in the ecosystems.

What first interested you in this field of research?

I still remember when I first looked at a meiofauna sample and I said to my friend “Why do you keep sand in these pots” haha. Since then, I have dedicated my research time to this group. Understanding the magnitude of biodiversity is one of the most important objectives of life scientists. Several groups have had their knowledge barriers being overcome such as the meiofauna. This is an important fraction of biodiversity in the earth, with a lot of functions in the environments such as beaches, rivers, mangroves, deep-sea, etc. Understanding how these communities function and how they contribute to global biodiversity is one of my objectives in life.

Can you briefly explain the research you presented at 18IMCO?

The aim of my  project was to develop a methodological tool based on machine learning to help in the identification of free-living marine nematode species. A total of 40 species belonging to the nematode genus Acantholaimus and 58 belonging to the genus Sabatieria were considered. For both genera, morphometric and presence/absence morphological  characters were considered. For testing the methodology compared K-nearest neighbor (KNN), Support Vector Machine (SVM), Random Forest (RF), and Gradient Boosting (GB) algorithms.  As result for both genera, RF was the most accurate in classifying the specimens into the correct species (94%), while K showed the worst
performance (17%). After the ensemble of RF and SVM, the accuracy rose to 99.7% for Sabatieria and 100% for Acantholaimus. These results showed that, in the presence of a morphometric table, the identification of marine nematodes could be fully automated, optimizing biodiversity and ecological studies as well as making species identification more accessible for non-taxonomists.

How will you continue to build on this research?

The next step is to present through this methodological tool to develop a digital application for use of this tool by other researchers. The application will be developed in order to facilitate both the entry of new data from other genera and the identification of new morphotypes by morphometry and morphology.

We are thinking of working with the genera of the Cyatholaimidae, Chromadoridae and Selachenematidae families. The choice of these families is because they have good reviews and species descriptions. The first step is to carry out a survey in the literature of the main morphometric and  morphological characters and categories used in the separation of species. The next step will be to include the described species and individuals each species  found in the literature in the matrix and then generate the identification model for the species through the algorithms of machine learning.

You may also like...