Search form

Vlaams Instituut voor de Zee

Contact details
Traineeship proposition
Voorstel stage-onderwerp 2012-2013
In het Vlaams Marien Datacentrum, een onderdeel van het VLIZ, beheren we verschillende soortenlijsten. Deze worden door taxonomen gecontroleerd en uitgebreid, zie lager voor de urls. Graag werken we in het kader hiervan een stage uit voor een student Bio-Informatica.
WoRMS : (World Reg. Mar. Species)
ERMs : (Eur. Reg. Mar. Species)
BeRMs : (Belgian Reg. Mar. Species)
Abstract Bachelor Project FBT 2019-2020The representativity of zooplankton samples in the Belgian part of the North Sea processed with the ZooSCAN

Without zooplankton, life on earth would be completely different. The central location in the ecosystem and the important functions in carbon storage make these organisms extremely important for all life on earth. Changes within the zooplankton population would lead to massive mortality in the ecosystems and would mean an extreme increase of carbon dioxide in the atmosphere. Climate change issues, overfishing and the construction of windmills are putting zooplankton species in danger. Rising ocean temperature and ocean acidification would mean the end of zooplankton.

Due to its high sensitivity and extremely important functions, zooplankton is seen as one of the most important indicators of climate change. This is why monthly samples are taken at various locations in the Belgian part of the North Sea by 'het Vlaams Instituut voor de Zee'. These locations are called ‘stations’. One sample is taken at each station to show the population at the sampling location. The extreme variations in densities between stations make it necessary to check whether these variations also occur at a smaller distance from each other. If these variations occur inside a station, the results of the last few years are inaccurate or even wrong, and a sample would be a 'lucky shot’. To check this, five replicates were taken at two different stations.

The samples were analyzed with the ZooSCAN, a revolutionary device to analyze zooplankton samples. Each replicate was analyzed five times in order to be able to identify variations by analyzing them. Using statistical data processing, the variations caused during the analysis can be compared with the variations between the replicates and between the different stations in the same month. In the ideal case, replicates are completely the same and the spread would be similar to the variation caused during the analysis.

The results indicate that the variation between the replicates in the large groups are somewhat higher than the variations observed during analysis. The higher variation has a natural cause and implies that the densities of these species fluctuate somewhat inside a station. For the smaller groups, the results are more encouraging, and the variations are no more significant than those observed during the analyses. This means that the observation of these species is done correctly. When the variations of the replicates are compared with the fluctuations between the different stations, the spread proves to be a lot smaller for the replicates.

It can be concluded that one sampling gives a good indication of the composition of the zooplankton population in the sampling site. Extra samples inside a station or extra analyses with the ZooSCAN can make the results more accurate.

Abstract advanced bachelor of bioinformatics 2019-2020Combining marine species occurrence data into standardized maps for analysing their distribution

A distribution map is being developed which combines the occurrence records of different databases to study the distribution of marine species and to also see the links between these databases.

The project is envisioned as follows. Using R, all the occurrence records are downloaded from the WoRMS, OBIS, EurOBIS, GBIF and IDIGBIO databases. These are then combined in to a MongoDB database, along with the shapefile of Marine Regions (MR) and Marine Ecoregions Of the World (MEOW). Each shapefile contains a list of polygons, which are linked to the occurrence records via their (latitude/longitude) coordinates. The records also contain a date/time value in most cases. Through an R Shiny app the user can choose the location, taxon and a date range as input. The R Shiny app will query the MongoDB database for the right records and polygons, which will then be plotted on a leaflet map as points and polygons. The infrastructure of this distribution map will be added to the Lifewatch R Shiny projects and to the WoRMS distribution map.

In this traineeship the download is still in progress. The code to download the records is written and is now being applied to download all the records. The MongoDB structure has also been determined.

Further steps of the project, which is the R Shiny app and integration into Lifewatch and WoRMS are still to be done.


Abstract advanced bachelor of bioinformatics 2018-2019: Automation of link updates between taxonomic and genetic databases
WoRMS ( is a website that consists of an authorative and comprehensive list of marine organisms. It is a perfect searchplace for scientists to find potential subjects to perform analysis on. One feature that helps with this search is the links that WoRMS offers to scientists to other scientific sites like the NCBI taxonomy browser (Genbank), European Bioinformatics Institute (EBI), Barcode of Life (BoLD) and many more. This project focusses on the automation of link updates of three different genetic databases in WoRMS using PHP scripting. Taxonomy browser links will be updated much quicker to show a more accurate overview of the number of nucleotides and proteins per organism. EBI links and Elixir MarRef links were added as new link providers. A general workflow was used to make all the links. A series of steps were performed on the downloaded files to prepare them for implementation in a database. when implemented in the database the links could be made. Future versions of the scripts may include more info depending on the user’s demand.
The purpose of my internship was to verify the ZooSCAN process and its optimization. It’s a fairly recent technology and a comprehensive check on its operation had not been done yet at the Flanders Marine Institute (VLIZ). Three crucial parts of the process have been thoroughly examined. Initially the focus was on sample preparation. With dilution series the ‘Motodo Plankton Splitter’ is tested. Verification on how many dilutions can be made and still have representative results. The system replicates have been tested in detail and extensively. The final step was to evaluate the learningsets which are used to classify zooplankton. This bachelor thesis can be used as a manual for this infrastructure and enable future users to quickly understand the ZooSCAN. The features are summarized.
Biodiversity research study is ecologically important because zooplankton is an essential part of the marine ecosystem in the North Sea. The ecological quality of these waters is determined inter alia by the concentrations of various species zooplankton. Therefor the zooplankton should be monitored, because it is the basic food of many marine animals, especially for higher trophic levels and specifically to fish. In addition, zooplankton is highly temperature-dependent and thus perfect to capture changes in climate.
The RV Simon Stevin sails on regular basis to collect zooplankton. Collected samples have to be sieved first to extract the medium. If necessary dilutions can be made. Afterwards, the samples are scanned with the ZooSCAN. Computer programs used for investigating are: ImageJ, VueScan and the statistical program Tanagra. The last step is the taxonomic classification for which the computer program Plankton Identifier is used.
Sample preparation has proven to be more efficient with a 200 µm sieve. Zooplankton should be separated from each other on the sample bed before scanning. If sand is present this can best be decanted with glass measuring cups.
The number of converted fragments after dilution should be around 2000. This enables quick and efficient work. If there are very low numbers of zooplankton in the validated groups, the entire process should be repeated. Alternatively two smaller fractions added together can obtain an appropriate dilution.
The existing VLIZ learningset works best if merged with the newly created set. The classification is than better and more time efficient. In the future, vignettes of less common taxa have to be added to the learningset. This will significantly improve the software.


Wandelaarkaai 7
8400 Oostende
+32 [0]59/34 21 30


Traineeship supervisor
T'Jampens Roeland
Traineeship supervisor
Bart Vanhoorne
+32 59 34 21 30
Jonas Mortelmans
Via Map