Events News Research CBS CBS Publications Bioinformatics
Staff Contact About Internal CBS CBS Other

Behavioral Phenomics

Integrative Systems Biology of Phenotype and Genotype


Group leader: Hanne Østergaard Jarmer
Member: Karin Marie Brandt Wolffhechel
Guest members: Agnieszka Sierakowska Juncker, Anders Gorm Pedersen, Aron Charles Eklund, Henrik Bjørn Nielsen, Johanne Ahrenfeldt, Ramneek Gupta, Simon Rasmussen, Thomas Sicheritz Pontén, Tune Hannes Pers, Ulrik Plesner Jacobsen

Catalog of previous student projects.


About the group

The core of the research in the Behavioral Phenomics group resides in finding the connections between human genotype and phenotype - that being any aspect of physical appearance, personality, intelligence, voice, behavior, physical strength, pathological psychiatric disorders, etc. Such insight will enable the prediction of the investigated phenotypes, to the extent that these can be explained by genetics. The main difference from our work to what has been done previously, starting with the famous pea experiments, is that we use an endophenotype-based approach to describe the parameters that we wish to investigate. We aim at not only identifying the genetics, but strive to build a method to predict, describe or even reconstruct the appearance of the given phenotype. Also, we do not limit ourselves to any specific type or source of data; we integrate data from all relevant sources.

Goals:
- Identify meaningful junctions between human genetics, behavior and psychology
- Predicting human appearance from genetics to enable face recognition

We are currently finishing up a Proof of Concept (PoC) study in which we link facial-recognition techniques, Genome-Wide Association Studies (GWAS) and classification methods to predict human facial features from genotype information.

logo
FullFace

Appearnetics - From DNA to Portrait

In this project we have established a strategic collaboration across two research areas that traditionally do not interact: image analysis and genome-wide association studies. The aim is to create a tool that converts the genetic information derived from DNA into a portrait of the person to whom the DNA belongs. In the above briefly mentioned PoC study ~600 icelandic individuals have been used as input, and two vectors for each of these have been extracted. The first describes the person's face - the face key, which is the outcome of the image analysis using methods classically employed in facial recognition software. The second vector is derived from a genetic analysis of the subject's genome, and contains the digital information about approximately 17 million genetic variants (Single Nucleotide Polymorphisms, SNPs) of the given subject. By association analyses, we are identifying the genetic constellations that determine the values of the face key, and hence enable the process of reverse engineering a portrait from DNA. For more information about the project, please, see: From DNA to Portrait.

This project is performed in close collaboration with Associate Professor Rasmus Reinhold Paulsen and PhD-student Jens Fagertun from DTU Informatics. The project is also supported by several both national and international entities.

Personality Trait Prediction

From existing data describing the personality and genotypes of many thousand Europeans we are planning to conduct a study methodologically similar to the above described appearnetics project.

Personality traits are psychological phenotypes, such as the well-described "Big Five": Openness_to_experience, Conscientiousness, Extraversion, Agreeableness and Neuroticism. As the name indicates, these have been shown to describe personality quite well. They have been characterized through a factor analysis performed using the psychometric measurements from many thousands of individuals. The psychometrics for a person is obtained through a large standardized questionaire commonly known as a personality test (such as the Revised NEO Personality Inventory test). Each of these 5 super traits are the sum of several sub-traits. As an example, in Neuroticism we find: Anxiety, impulsiveness, depression, hostility, self-consciousness and vulnerability to stress.

Personality traits have via twin studies been shown to be hereditary to a large extent, and on average roughly half of the variation can be explained from genetics. Genome-Wide Association Studies have revealed a number of causative SNPs explaining a minor fraction of some of the "Big Five", but so far nobody has attempted to predict personality from genotype data.

So, why would we want to predict personality? Why do this, when the personlity tests are already fulfilling the task? The answer is a myriad of reasons. Most importantly, the work would contribute high-value basic research in the genetics of human personality, facilitating the understanding of human behavior and the characterization of borderline or pathological psychiatric disorders. Also, it is not always possible to perform a personality test. The person of interest could be dead or missing (leaving behind DNA). The person could be a very young child or a psychiatric patient unable to do the test. A DNA-based prediction of personality would also help solve the problem that a self-report qustionaire may be filled out according to what the given person believes are the "correct" answers seen in the light of the purpose for the test, and in this way disguise the true personality - often even unintentionally.
FullFace FullFace

Teaching

Teaching is of high priority to the group. We see it as a privilege to be part of the education of engineers at a high academic level. Our goal is to keep developing our teaching on all levels. For this process to be successful any feedback from both colleagues and students will be most appreciated. We teach the course: DNA Microarray Analysis (DTU course #27612) every January.

Project work in Behavioral Phenomics

Students who would like to write their bachelor thesis in the group are advised to take an introductory course in bioinformatics such as Introduction to Bioinformatics - #27611, and a programming course such as Perl and Unix for Bioinformaticians - #27619. Similarly master students are in addition encouraged to follow the 3-weeks course in January: DNA Microarray analysis - #27612


Full list of CBS publications