Logo: to the web site of Uppsala University

Electroencephalogram (EEG) interpretation plays a critical role in the clinical assessment of neurological conditions, most notably epilepsy. However, EEG recordings are typically analyzed manually by highly specialized and heavily trained personnel. Moreover, the low rate of capturing abnormal events during the procedure makes interpretation time-consuming, resource-hungry, and overall an expensive process. Automatic detection offers the potential to improve the quality of patient care by shortening the time to diagnosis, managing big data and optimizing the allocation of human resources towards precision medicine. Here, we present MindReader, a novel unsupervised machine-learning method comprised of the interplay between an autoencoder network, a hidden Markov model (HMM), and a generative component: after dividing the signal into overlapping frames and performing a fast Fourier transform, MindReader trains an autoencoder neural network for dimensionality reduction and compact representation of different frequency patterns for each frame. Next, we processed the temporal patterns using a HMM, while a third and generative component hypothesized and characterized the different phases that were then fed back to the HMM. MindReader then automatically generates labels that the physician can interpret as pathological and non-pathological phases, thus effectively reducing the search space for trained personnel. We evaluated MindReader’s predictive performance on 686 recordings, encompassing more than 980 h from the publicly available Physionet database. Compared to manual annotations, MindReader identified 197 of 198 epileptic events (99.45%), and is, as such, a highly sensitive method, which is a prerequisite for clinical use.

Background: Electroencephalogram (EEG) interpretation plays a critical role in the clinical assessment of neurological conditions, including epilepsy. Manual analysis requires highly specialized and heavily trained personnel. Moreover, the rate of capturing abnormal events makes interpretation time-consuming, resource-hungry, and, overall, an expensive process.

Automatic detection offers the potential to improve the quality of patient care by shortening the time to diagnosis, managing big data, and optimizing the allocation of human resources.

Findings: We present MindReader, an unsupervised method for EEG signals. First, MindReader processes the signal through an autoencoder in order to detect EEG abnormalities. Next, patterns are hypothesized by a Hidden Markov Model. Our algorithm automatically generates labels for non-pathological phases, thus reducing the search space for trained personnel.

Conclusions: MindReader is effective in detecting EEG abnormalities in focal and generalized epilepsy.

The organization of living beings is complex. Science uses modeling in order to gain a deeper understanding, and to be able to manipulate the processes of living organisms. To this purpose, I used and developed computational tools to investigate and model different relevant biological phenomena. 

In paper I, I utilized whole-genome data from wild and domesticated European rabbit (Oryctolagus cuniculus sp.) populations to identify segregating insertions of endogenous retroviruses and compare their variation along the host phylogeny and domestication history. The results from this study highlight the importance of genomic modeling beyond reference organisms and reference individuals, and provide deep insights regarding strategies for variant analyses in host population comparative genomics. In paper IV, I studied the process of exaptation of foreign genetic elements at broad-scale by observing the presence and characteristics of retroviral env gene, syncytin, across vertebrates. I searched a library of more than 150 chromosome-length assemblies covering 17 taxonomical orders for syncytin homologs, where I identified and syntenically aligned over 300 loci insertions, including not previously known insertions. Additionally, three-dimensional structures of the recovered sequences were predicted using AlphaFold2. Phylogenomics analyses suggest a complex dynamic of multiple retroviral insertions at different time points with sequence conservation specific to clades that share a similar histo-physiological placental type.

In paper II, I expanded the scope to encompass translational medicine by developing an unsupervised machine learning methodology for detecting anomalies in biomedical signals, MindReader, which I applied primarily to electroencephalogram. In paper III, I developed a hidden Markov model implementation that includes a hypothesis generator for stream time-domain signals, which is used as a dependency for paper II. The work in this thesis substantiates that a combination of biological knowledge, cutting-edge technology, and robust algorithmic design constitute the primordial factors for scientific advancement.

Although recent advances in sequencing and computational analyses have facilitated use of endogenous retroviruses (ERVs) for deciphering coevolution among retroviruses and their hosts, sampling effects from different host populations present major challenges. Here we utilize available whole-genome data from wild and domesticated European rabbit (Oryctolagus cuniculus sp.) populations, sequenced as DNA pools by paired-end Illumina technology, for identifying segregating reference as well as nonreference ERV loci, to reveal their variation along the host phylogeny and domestication history. To produce new viruses, retroviruses must insert a proviral DNA copy into the host nuclear DNA. Occasional proviral insertions into the host germline have been passed down through generations as inherited ERVs during millions of years. These ERVs represent retroviruses that were active at the time of infection and thus present a remarkable record of historical virus–host associations. To examine segregating ERVs in host populations, we apply a reference library search strategy for anchoring ERV-associated short-sequence read pairs from pooled whole-genome sequences to reference genome assembly positions. We show that most ERVs segregate along host phylogeny but also uncover radiation of some ERVs, identified as segregating loci among wild and domestic rabbits. The study targets pertinent issues regarding genome sampling when examining virus–host evolution from the genomic ERV record and offers improved scope regarding common strategies for single-nucleotide variant analyses in host population comparative genomics.

Syncytin is a fossil protein exapted from retroviruses that fulfills a pivotal role during trophoblast implantation and placental metabolite exchange. However, little is yet known about the distribution of syncytin across vertebrates. Here, we searched a library of more than 150 high-quality assemblies across 17 taxonomical orders for syncytin homologs. We identified and syntenically aligned over 300 loci insertions, including not previously known insertions. Additionally, we predicted the tridimensional structures of the recover sequences using AlphaFold2. Sequence conservation and phylogenomics analyses suggest a complex dynamic of multiple retroviral insertions at different time points with sequence conservation specific to clades that share a similar histo-physiological placental type. This research has widened our knowledge about the physiology of placentation through a better understanding of the evolutionary role of syncytin.

Transposable elements represent a substantial proportion of eukaryotic genomes, where they can disrupt or enhance gene expression on the host. However, identification at population scale where often short sequencing signals are available is challenging. Current approaches rely on parsing sequence alignment files looking for anomalies on read length, read orientation and read depth, but they are often slow and complicated to install. Here, we present the Chapulin, a portable cross-platform, open-sourced Rust application for structural variant identification and characterization, including transposable elements. By using concurrent computing and native execution, Chapulin identifies a large fraction of mobile element insertions while outperforming existing transposable element tools. Chapulin was designed to be versatile and robust, in order to accommodate the demands of current data, such as population-scale studies or clinical samples