Javascript verkar inte påslaget? - Vissa delar av Lunds universitets webbplats fungerar inte optimalt utan javascript, kontrollera din webbläsares inställningar.
Du är här


Projektförslag för kandidatexamensarbete 15 hp Vt2020, 24 mars - 5 juni

Two projects in Botany:

  • Analyzing (mostly already existing) data on relationships between plant species richness/diversity and landscape parameters in Skåne.. The basis would be already existing species lists and species counts for 2.5*2.5 km grid-squares which would be interesting to relate to different kinds of explanatory parameters that may be extracted from maps using more or less advanced GIS technologies (depending on the skills and interests of the student) to show how plant biodiversity depend on e.g. climatic, geological and land-use parameters.
  • Bryophyte diversity on church-yards in Skåne using already existing bryophyte data. Such a project could then also include either some simple remote sensing based on aerial photographs and/or some field-work at the actual church-yards (then I would supply the bryophyte data while the student may herself collect the ”explanatory variables”).

Supervisor: Torbjörn Tyler


Pollen analysis using deep learning

Deep learning is a group of methods in artificial intelligence, also knowns as convolutional neural networks (CNN). They are particularly useful for tasks such as image recognition, and there are a number of open-source CNNs that have been trained to identify up to 1000 classes of images, such as cats, dogs, cars, and ice cones. These models can be retrained to identify other classes of images, given sufficient training images. 
Pollen analysis is the identification and quantification of pollen from different flower species. This is an important tool when studying the plant-pollinator interactions in ecological research, and when studying and quantifying pollination as an ecosystem service. However, also in fields as different as paleoecology, allergen quantification, and quality assessment of honey. To date, most pollen analysis is done manually by studying samples in a microscope, which is both difficult and time consuming. 
Currently, we are developing pollen analysis methods based on deep learning techniques. We have models trained to identify ca 20 different pollen classes (species, or very similar looking species) that work well, and are sufficient for some ongoing tasks. However, we also have images of over 200 additional species in a pollen library, waiting to be used. This project would be about preparing the training datasets, train new models, and analyze their performance. 
There are many different aspects to study regarding the performance of the CNNs, when applied to a new field like this. For example, what features does the AI use to identify a particular species of pollen? Are these the same features as a human expert would use, or a different set of features? The similarities or differences between pollen species, according to the CNN can be measured, and analyzed in a cladogram. Will such a cladogram resemble a taxonomic cladogram, or will it be different? Will a single model with all different species be the most efficient, or will it better to have several models for similar species? These, and other similar questions, will be the focus of this project. 
The project is suitable for two students, who can collaborate on most parts, but select slightly different foci for their analyses and reports. Most of the work with preparing the training datasets can be done on any computer, and at home if necessary. The actual training of the CNNs is very computer intensive, and must be done at one particular computer at the university, with assistance from the supervisor. Analyses of the data will be relatively straightforward.

Supervisor: Ola Olsson


Sex identification of white storks based on skull and bill measurements

Male and female white storks (Ciconia ciconia) look exactly the same, but on average males are slightly larger. The storks are reintroduced in Sweden, in an ongoing project that started in 1989 (, and in the captive breeding program it is important to know which bird is a female, and which is a male. It is possible to use molecular analyses to identify the sex of the storks, but this is both costly and takes time. As alternative, there are preliminary data that indicates that the male has long bill, but this has not been properly confirmed. Measuring the bill is quick, and could potentially give an immediate sex identification.
This project will be about using existing data of living storks’ body measurements, and complement it with new measurements of skeletons stored at the zoological museum. Furthermore, many of the individuals in the museum were measured while alive, which provides the opportunity to assess how accurate the measurement of the living bird is to measure the skull and bill.
Measuring the skeletons will be done in the zoological museum’s collections at Arkivcentrum Syd. Data analysis and writing can be done anywhere.

Supervisor: Ola Olsson



Lobula plate neuron analysis: 3D reconstruction and electrophysiology analysis 

We have injected fluorescent dye in to a hoverfly (Eristalis tenax) lobula plate neuron and also recorded intracellular activity while showing visual stimulus on a screen in front of the animal. This neuron is used as a control neuron in 
another project but has a nice dye injection and interesting response properties and thus deserves more analysis. It is a type of neuron that responds large and small moving objects (prefers large) from the right visual field (electrode in left side of brain). The project would involve:

  • 3D reconstruction of the neuron from a confocal image stack
  • Analysis of electrophysiological responses to visual stimulus
  • Review of literature on lobula complex neurons to improve an existing preliminary characterization list
  • Estimate neuron subtype based on the improved characterization list
  • Analyze dendritic tree and compare with other neuronal reconstructions from e.g.
  • Speculate in what types of visual processing the neuron could be involved in such as object motion detection,
  • short term memory, attention and optic flow.


Bo Bekkouche and David O’Carroll


Morphological comparison of object motion detector neurons

Recently a large part of the drosophila melanogaster (fruit fly) brain was imaged using a technique called focused ion 
beam and scanning electron microscopy (FIB-SEM). This dataset (called hemibrain) has enabled thousands of neurons to 
be reconstructed including several previously neuron subtypes that have not been analyzed and compared 
morphologically. Our lab is interested in object motion detecting neurons and we are focusing mainly on hoverflies and 
dragonflies with the fruit fly and blowfly as comparison species. This project would involve:

  • Comparison of 3D reconstructions of object motion detector neurons and analyze network structure of dendritic trees
  • Potentially perform cluster analysis or use machine learning to train a classifier of object motion detector 
  • subtypes
  • Comparison with hoverfly/dragonfly object motion detector neurons
  • Visualize network of dendritic trees and speculate how they would connect in a model
  • Potentially design a morphologically detailed network model and do some basic tests of it 
  • Potentially connect model to a bioinspired small object motion detector model
  • Potentially try to integrate a graded synapse between the neurons in the network


Bo Bekkouche and David O’Carroll


Simulating facilitation in a spiking neural network

In our lab we are actively designing various types of computer simulation models of the insect visual system. One of the things we want to simulate is a type of short term memory of object motion that we observe in so called small target motion detector neurons in the dragonfly (and potentially Hoverfly). One type of model which we have not yet spent much time on modeling is the leaky integrate and fire model (IAF or LIF) which is a simple conductance based model that fires action potentials (using e.g. NEST simulator) and is often used to create large neural networks due to its relatively low computational processing demand. It is however not clear if this type of model is capable of simulating facilitation as observed in the in vivo data from dragonflies, and it is especially questionable if it can simulate the traveling facilitation wave. This project would involve:

  • Reviewing the types of synaptic short term memory models for LIF or LIF-like neurons that exist and could be 
  • utilized for simulating facilitation
  • Creating a network model (in NEST simulator, using python) using a candidate short term potentiation synapse, and simulate facilitation.
  • Investigating if any of the synaptic models could create a traveling facilitation wave
  • Investigating and trying to find graded synapse models in NEST simulator or similar simulator and test if it can be
  • used to simulate a traveling wave


Bo Bekkouche and David O’Carroll



Building for bees - Using bee hotels and citizen science to learn more about wild bees

In this project, you will use citizen science data gathered by Naturskyddsföreningen to answer questions about cavity-nesting solitary bees using bee-hotels. For example, which types of environments do these bees use most, how important are gardens in urban areas, compared to rural sites, and does the amount of flowers in the garden affect the number of nesting bees?
Pollinating insects, such as wild bees, are declining worldwide, which may reduce pollination of both wild plants and crops. The main reasons for the declines are losses of habitat and flowering food resources, but a lack of nesting habitats may also limit populations. In recent years, urban areas have been proposed to offer refuges for pollinators, at least in comparison to farmland landscapes, and residential gardens may be particularly good habitat. 
The amount of flowers in gardens can be increased by turning lawns into flowering meadows, or choosing to plant nectar and pollen-rich plants in flowerbeds. The amount of nest sites for cavity-nesting solitary bees can be increased through so-called bee hotels (for example drilled holes in wood or bunches of hollow bamboo sticks). However, the benefit of bee hotels has been questioned. In 2018, the Swedish Society for Nature Conservation (Naturskyddsföreningen) started a campaign to increase the awareness of the decline of wild bees and urge people to increase the amount of flowers and put bee hotels in their gardens. They also asked people to report the occupancy in the hotels at the end of the season, along with information on environmental data such as the amount of flowers in the surrounding. In 2019, 10.000 people participated and ca 2000 reported data back to us.

You can use existing data (2018 and 2019) or make a literature review of bee hotels as a method to benefit bees by compiling results from published papers. Because this is a citizen science project, it is important to communicate the results back to the participants.

Methods and skills needed: Compiling scientific literature on bee hotels. You need to be able to use statistical software (R or SPSS).


Anna Persson, Centre for Environment and Climate research (CEC), Email: anna [dot] persson [at] cec [dot] lu [dot] se


Jackdaw flight – wing motion across speed

The flight of animals has been studied for centuries, but it is only in rather recent time that we are able to quantitatively study the details of the kinematics and aerodynamics of the animals in flight. The technical development now allows us to determine the cost of flight and how it varies with flight speed.

In this bachelor project you will be involved in a study focused on the flight of the Jackdaw. The problem – and this is where you come in – is that no one really knows how the animal controls its force production and the power required to fly.

The goal for the study is to be able to find out how the birds control the forces to accomplish efficient flight, by linking their wing movement patterns – the kinematics - to the aerodynamic forces generated and the power required to fly.

You will be analysing existing data in the form of kinematic high-speed movies of a Jackdaw filmed flying in the Lund University Wind Tunnel at various flight speeds. You will be tracking the motion of the animals from the videos and reconstruct 3D coordinates of how the wings move over time throughout the flights. From the data you will estimate aerodynamically relevant parameters, such as the amplitude of the wingbeat and the angle the air meets the wing. You will be able to use existing corresponding power measurements that have been acquired through analysis of the aerodynamics, to compare the kinematic results with.


Christoffer Johansson, Christoffer [dot] Johansson [at] biol [dot] lu [dot] se


Understanding molecular processes in barley

In our research group we are trying to understand molecular process in plants related to design of plant architecture, time to flowering, chlorophyll biosynthesis and wax biosynthesis. We often use historic barley (Hordeum vulgare L.) mutants which were isolated more than 50 years ago and analyze them with modern molecular techniques and biochemical methods. Thus, our research spans from DNA sequencing (both Sanger sequencing and next-generation sequencing) to characterization of proteins. Our projects address fundamental scientific questions but is also of applied interest for plant breeding. 

The following three recent publications reflect what we are doing at present. Bachelor or master students participated in all three publications and ended up as co-authors.

•    Braumann, I., W. Urban, A. Preuβ, C. Dockter, S. Zakhrabekova and M. Hansson. 2018. Semi-dwarf barley (Hordeum vulgare L.) brh2 and ari-l mutants are deficient in a U-box E3 ubiquitin ligase. Plant Growth Regul. 86: 223-234.

•    Schneider, L. M., et al. 2016. The Cer-cqu gene cluster determines three key players in a -diketone synthase polyketide pathway synthesizing aliphatics in epicuticular waxes. J. Exp. Bot. 67: 2715-2730.

•    Dockter, C., et al. 2014. Induced variations in brassinosteroid genes define barley height and sturdiness, and expand the “Green Revolution” genetic toolkit. Plant Physiol. 166: 1912-1927.

We have a lot of ideas about possible bachelor and master projects, so please come and discuss with us. We want to hear what you want to learn and that help us to design a suitable project.


Mats Hansson, mats [dot] hansson [at] biol [dot] lu [dot] se)'.



Does sex-specific selection change mating behaviour in a hermaphrodite?


We have attempted to mimic the early stages of sex chromosome evolution by subjecting the simultaneous hermaphrodite flatworm Macrostomum lignano to sex-limited selection using a GFP marker as a sex-determining locus. After 21-22 generations of sex-limited experimental evolution, we found that female-selected worms had increased egg production and decreased sperm competitiveness compared to controls and male-selected worms. We also found changes in morphology (body size, testes size, and ovary size) consistent with the observed fitness differences. Now after more than 40 generations we are interested to find out if mating behaviour has also changed. It has been suggested that, unless sperm limitation is common, hermaphrodites that mate often probably do so because they want to give away sperm rather than receive it. In addition, these worms have evolved a suck behaviour, where the worm moves its head onto genital opening after mating and apparently tries to get rid of partner’s sperm and/or other substances transferred within. Thus, we expect that female-selected lines will mate less frequently as well as perform suck behaviour less often than male-selected lines.
To measure mating behaviour, we have video recorded the worms and now number and length of matings and suck behaviour will be observed. So far preliminary analyses show that male-selected worms indeed mate more and that worms mated with these male-selected worms tend to perform suck behaviour more often. Now we want to check if the length of the mating (and sucking) has changed and if there are some interactions with frequencies of these behaviours in selected worms (e.g. if worms which mate longer, mate less frequently). The student working on this project will therefore learn how to collect, analyse and present behavioural data, as well as the theory behind the research project.


Aivars Cirulis, MSc (aivars [dot] cirulis [at] biol [dot] lu [dot] se)
Jessica Abbott, PhD (jessica [dot] abbott [at] biol [dot] lu [dot] se)


Why are hummingbirds so diverse in the Andes? 

Do you want to contribute to a better understanding of the unique diversity that is associated with, for example, pant-hummingbird diversity hotspot in the Andes? This is possible through a Bachelor project that aims to analyze the ecological and evolutionary processes that affect the structure and dynamics of biological communities that are distributed along environmental gradients (e.g. mountain ranges). 

The Masters project will be part of a larger research program that aims to enable i) a better understanding of observed diversity gradients, ii) quantification of eco-evolutionary processes that structures plant-hummingbird data and iii) prediction of biodiversity in a changing environment. Given this broad context, the student will have the opportunity to plan their own project and work with available simulation models of how diversity is assembled and maintained. The student may, for example, choose to simulate the assembly of theoretical communities and contrast the structure of communities that are distributed along environmental gradients with different slopes.
The project will involve quantitative methods and tools. We are thus looking for highly motivated students with a keen interest in biodiversity drivers as well as good computer skills and interest in mathematical modeling and computer programing. 


Mikael Pontarp
Forskare, Principal Investigator
mikael [dot] pontarp [at] biol [dot] lu [dot] se 


Farmland intensification and pest dynamics

A dilemma in modern agriculture is how to maintain or increase production, while minimizing the negative effects of increased field area, reduced crop diversity and increased pesticide use. One way to do so is to facilitate the abundance and diversity of ecosystem service-providing organisms, such as natural enemies of pests. This Bachelor project builds on available simulation models of biological control and it is part of a larger research program which aim to: I) develop a trait-based eco-evolutionary model of pest-predator dynamics in a heterogeneous agricultural landscape. II) Parameterize the model with data on, for example, functional traits on soil invertebrates and Carabids. III) Analyze model output in order to improve our understanding and prediction of observed pest-predator dynamics in Swedish and European agriculture. Given this broad context, the student will have the opportunity to plan their own project and work with available simulation models. The student may, for example, choose to simulate the assembly of a theoretical pest-predator community under different scenarios of landscape heterogeneity.
The project will involve quantitative methods and statistical tools. We are thus looking for highly motivated students with a keen interest in biology as well as good computer skills and interest in mathematical modeling and computer programing.


Mikael Pontarp
Forskare, Principal Investigator
mikael [dot] pontarp [at] biol [dot] lu [dot] se 


Who is staying, who is moving? Territory settlement in partially migratory Blackbirds

In partially migratory populations, some individuals migrate to overwinter somewhere else, whereas other individuals in the same population stay resident year-round. Blackbirds (Koltrastar på svenska) are a typical partial migrant. Last year I equipped 57 Blackbirds in Lund with radio-transmitters to determine which individuals migrated and which remained resident. The start of the Bachelor course coincides with the start of the reproductive period and birds will settle in their territories. Theory predicts that residents should have an advantage over migrants and take the best territories. If that is true, one would expect residents to maintain their territories from previous years more often, while migrants might be forced to move into unoccupied territories. To test this, the students will use data from an automated state-of-the-art telemetry system to determine which individuals have been migrants and residents in winter. Then they will use a hand-held radio-transmitter and locate the territories of the different birds and compare them with their settlement in the last breeding season. If the students are interested, the can even link this to physiological data that are available. 

Would be really cool to get 1-3 students for this project!

Arne Hegemann
+46 46 222 37 81, +46 72 996 58 03
arne [dot] hegemann [at] biol [dot] lu [dot] se 


”Hur håller sig världens nordligaste fågel varm?”

På den högarktiska ögruppen Svalbard (78-81°N) lever världens nordligaste stannfågel, svalbardripan (Lagopus muta hyperborea). Så långt norrut präglas livet av stor säsongsvariation, med 5 månader polarnatt under vintern och 5 månader midnattssol under vår och sommar, och det är ofta omkring 60°C kallare på vintern än på sommaren. I den här miljön måste svalbardripan hitta föda året om, vilket kan vara nog så svårt eftersom den endast lever av lågvuxen tundravegetation som är täckt av snö och is under stora delar av året. Den här extrema miljön har selekterat för utveckling av flera anpassningar för att spara energi och motstå långa perioder utan mat på vintern, särskilt en sänkning av energiförbrukning via lägre aktivitet och en anmärkningsvärd pålagring av fett som gör att kroppsvikten fördubblas mellan sommar och vinter. Under sensommaren ruggar svalbardripan dessutom från sin brungråa sommardräkt till en helt vit vinterdräkt som är isolerar fågeln mycket bättre när det är kallt. Vi har sett att detta leder till avsevärt lägre värmeförlust både i laborativa och fältbaserade studier. Vad vi inte vet är dock hur mikro- och makrostrukturen i de enskilda fjädrarna, och i dräkten som helhet, ändras när sommar blir vinter. I det här projektet kommer du att studera just dessa frågor, för att ta reda på vilka morfologiska förändringar som gör det möjligt att svalbardripan nästan inte förlorar någon värme alls under vintern – en unik anpassning bland världens fåglar.


Andreas Nord ,
Phone: +46 704 953 262
E-mail: Andreas [dot] Nord [at] biol [dot] lu [dot] se



Utveckling av fågelrepellent baserat på beteendeprinciper för att minska skador i ekologisk majsodling

Obs, detta projektet kan påbörjas först i maj!

Fröpredation av fåglar han orsaka betydliga skador på jordbruksgrödor, särskilt om fälten angrips vid sådd. Fåglar är utbredda i jordbrukslandskap och på grund av deras höga rörlighet och intelligens, kan de snabbt och effektivt lokalisera födökällor, inklusive grödor. Det finns i nuläget flera sätt att begränsa fågelskador, t.ex. betning med bekämpningsmedel (vilket är skadligt för miljön) och olika skrämseltekniker (vilket är arbetsamt och kostsamt). Nya metoder krävs dock för att minska användning av bekämpningsmedel i jordbruket och möjliggöra ekologisk odling av grödor som majs. En möjlig strategi är att utnyttja kunskap om fåglars sinnen och beteende för att utveckla repellent som är både miljövänliga och har låg risk för att fåglar habituerar till dem.
Målet med detta projekt är att utveckla en ny metod som kan minska fågelskador på ekologisk majsodling i samarbete med Hushållningssällskapet. Metoden kommer att baseras på fåglars beteendemönster, som neofobi och associativ inlärning. Fåglar kommer att erbjudas majsfrön som luktar/smakar illa och har potentiellt avskräckande färger, vilket bör resultera i att fåglar uppfattar dessa som oätliga. Vi ska testa effektiviteten av flera fröbehandlingar i majsfält i Skåne, och även besvara följande frågor: Vilka fågelarter är de huvudsakliga fröpredatorerna? Vad är deras födösöksbeteende och aktivitetsmönster? Hur påverkas skador i fält av fågelförekomst i närliggande landskapet? 
Insamling av data: Fågelinventering i närheten av försöksfält före experimentet. Bildanalys från åtelkameror i försöksfält – arter, födösöksbeteende, hur många frön som konsumeras, etc. Uppskatta antalet överlevande majsplantor och analysera detta i relation till fröbehandling och fågelförekomst. Detta projekt passar en till två studenter (beroende på omfattning av kandidatarbete, 15 eller 30 poäng), med avgränsning av fokusområden inom 1) fågelbeteende i relation till fröbehandling och 2) praktisk tillämpning med avseende på skadenivå på majsplantor. Planering, fältarbete och datainsamling kan göras i samarbete om två studenter är involverade. Körkort krävs för genomförande av projektet. 
Föreslagen tidsplan för fältarbete och rapport
1, 5 veckor för fågelinventering, försöksfält och omgivande landskap (11 – 29 maj)
1, 5 veckor för video/bildanalys (11 – 29 maj)
1 vecka för uppskattning av skador på majsplantor (1 – 5 juni)
1, 5 veckor för dataanalys och skrivande (8 – 17 juni)
Föreslagen litteratur:
Clark, L. 1998. Review of bird repellents. Proceedings of the Eighteenth Vertebrate Pest Conference (1998). 6. Proc. 18th Vertebr. Pest Conf. (R.O. Baker & A.C. Crabb, Eds.) Published at Univ. of Calif., Davis

Mason, R.J. and Clark, L. 1992. Nonlethal repellents: the development of cost-effective, practical solutions to agricultural and industrial problems. Proceedings of the Fifteenth Vertebrate Pest Conference 1992. 51.

Månsson, J. Skador på gröda. Sveriges lantbruksuniversitet. Senaste åtkomst 2018-10-02

John Skelhorn, J. and Rowe, C.2010. Birds learn to use distastefulness as a signal of toxicity. Proc. R. Soc. B 277, 1729–1734, doi:10.1098/rspb.2009.2092
Marples, N. M. and Roper, T. J. 1996. Effects of novel colour and smell on the response of naive chicks towards food and water. Anim. Behav. 51, 1417–1424


Maria von Post & Diana Rubene


Application of behavioural principles in reducing corn seed predation by birds


Obs, detta projektet kan påbörjas först i maj!
Birds predating on crop seeds can cause substantial losses to agriculture, especially at the planting stage. Birds are widespread and abundant in agricultural landscapes, and due to their high mobility and intelligence, they are able to quickly locate and assess preferred food resources, including crops. Current strategies for managing crop damage by birds include chemical treatment of seeds (which spreads toxins into surrounding environment), as well as various scaring methods (which are laborious and costly). In order to reduce pesticide use in agriculture and facilitate organic crop production, new methods are needed. One possible strategy is to exploit bird sensory systems and behavioural patterns to develop repellents that are both environmentally friendly and with a low risk of birds habituating to them.
This project aims at developing a method for reducing bird damage to organic corn production, in collaboration with Hushållningssällskapet. In doing this, we will exploit behavioural principles like avoidance learning and neophobia. We will present the birds with corn seeds that are baited with potentially aversive colours and ill tasting, in order to reduce their perceived palatability to birds. Several seed treatments will be tested in corn fields in Scania and, in addition to evaluating seed treatment effectiveness, we will answer the following questions: Which bird species are the main seed predators on corn? What are their diurnal foraging and activity patterns? What is their foraging behavior in the field? How does abundance of these bird species in the surrounding landscape affect the seed predation pressure in the field?
Data collection: Bird survey in surrounding of field sites prior to experiment. Image/video analyses from wildlife cameras in study fields – species identity, foraging activity patterns, how many seeds consumed, etc. Estimation of number of surviving corn plants and relating this to seed treatment and bird abundance. This project is suitable for two students, with divided focal areas on birds behavior in relation to treatment and practical applicability in relation to corn plant results. Planning, field work and data collection would be a joint effort. Drivers license is a requirement for the project.
1, 5 weeks of bird surveying, both field experiments and wider landscape (11 – 29 May)
1, 5 weeks video/image analysis (11 – 29 May)
1 week estimation of corn plant damage (1 – 5 June)
1, 5 weeks for data analysis and writing up the results (8 – 17 June)
Suggested reading list:
Clark, L. 1998. Review of bird repellents. Proceedings of the Eighteenth Vertebrate Pest Conference (1998). 6. Proc. 18th Vertebr. Pest Conf. (R.O. Baker & A.C. Crabb, Eds.) Published at Univ. of Calif., Davis

Mason, R.J. and Clark, L. 1992. Nonlethal repellents: the development of cost-effective, practical solutions to agricultural and industrial problems. Proceedings of the Fifteenth Vertebrate Pest Conference 1992. 51.

Månsson, J. Skador på gröda. Sveriges lantbruksuniversitet. Senaste åtkomst 2018-10-02

John Skelhorn, J. and Rowe, C.2010. Birds learn to use distastefulness as a signal of toxicity. Proc. R. Soc. B 277, 1729–1734, doi:10.1098/rspb.2009.2092
Marples, N. M. and Roper, T. J. 1996. Effects of novel colour and smell on the response of naive chicks towards food and water. Anim. Behav. 51, 1417–1424


Maria von Post & Diana Rubene