| 1. |
|
|
| 2. |
- Almagro Armenteros, José Juan, et al.
(författare)
-
SignalP 5.0 improves signal peptide predictions using deep neural networks
- 2019
-
Ingår i: Nature Biotechnology. - : Springer Science and Business Media LLC. - 1087-0156 .- 1546-1696. ; 37:4, s. 420-423
-
Tidskriftsartikel (refereegranskat)abstract
- Signal peptides (SPs) are short amino acid sequences in the amino terminus of many newly synthesized proteins that target proteins into, or across, membranes. Bioinformatic tools can predict SPs from amino acid sequences, but most cannot distinguish between various types of signal peptides. We present a deep neural network-based approach that improves SP prediction across all domains of life and distinguishes between three types of prokaryotic SPs.
|
|
| 3. |
- Armenteros, Jose Juan Almagro, et al.
(författare)
-
Detecting sequence signals in targeting peptides using deep learning
- 2019
-
Ingår i: Life Science Alliance. - : LIFE SCIENCE ALLIANCE LLC. - 2575-1077. ; 2:5
-
Tidskriftsartikel (refereegranskat)abstract
- In bioinformatics, machine learning methods have been used to predict features embedded in the sequences. In contrast to what is generally assumed, machine learning approaches can also provide new insights into the underlying biology. Here, we demonstrate this by presenting TargetP 2.0, a novel state-of-the-art method to identify N-terminal sorting signals, which direct proteins to the secretory pathway, mitochondria, and chloroplasts or other plastids. By examining the strongest signals from the attention layer in the network, we find that the second residue in the protein, that is, the one following the initial methionine, has a strong influence on the classification. We observe that two-thirds of chloroplast and thylakoid transit peptides have an alanine in position 2, compared with 20% in other plant proteins. We also note that in fungi and single-celled eukaryotes, less than 30% of the targeting peptides have an amino acid that allows the removal of the N-terminal methionine compared with 60% for the proteins without targeting peptide. The importance of this feature for predictions has not been highlighted before.
|
|
| 4. |
- Bagger, Frederik Otzen, et al.
(författare)
-
HemaExplorer: a database of mRNA expression profiles in normal and malignant haematopoiesis.
- 2013
-
Ingår i: Nucleic Acids Research. - : Oxford University Press (OUP). - 1362-4962 .- 0305-1048. ; 41:D1, s. 1034-1039
-
Tidskriftsartikel (refereegranskat)abstract
- The HemaExplorer (http://servers.binf.ku.dk/hemaexplorer) is a curated database of processed mRNA Gene expression profiles (GEPs) that provides an easy display of gene expression in haematopoietic cells. HemaExplorer contains GEPs derived from mouse/human haematopoietic stem and progenitor cells as well as from more differentiated cell types. Moreover, data from distinct subtypes of human acute myeloid leukemia is included in the database allowing researchers to directly compare gene expression of leukemic cells with those of their closest normal counterpart. Normalization and batch correction lead to full integrity of the data in the database. The HemaExplorer has comprehensive visualization interface that can make it useful as a daily tool for biologists and cancer researchers to assess the expression patterns of genes encountered in research or literature. HemaExplorer is relevant for all research within the fields of leukemia, immunology, cell differentiation and the biology of the haematopoietic system.
|
|
| 5. |
|
|
| 6. |
- Bergenstråhle, Ludvig
(författare)
-
Computational Models of Spatial Transcriptomes
- 2024
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Spatial biology is a rapidly growing field that has seen tremendous progress over the last decade. We are now able to measure how the morphology, genome, transcriptome, and proteome of a tissue vary across space. Datasets generated by spatial technologies reflect the complexity of the systems they measure: They are multi-modal, high-dimensional, and layer an intricate web of dependencies between biological compartments at different length scales. To add to this complexity, measurements are often sparse and noisy, obfuscating the underlying biological signal and making the data difficult to interpret. In this thesis, we describe how data from spatial biology experiments can be analyzed with methods from deep learning and generative modeling to accelerate biological discovery. The thesis is divided into two parts. The first part provides an introduction to the fields of deep learning and spatial biology, and how the two can be combined to model spatial biology data. The second part consists of four papers describing methods that we have developed for this purpose. Paper I presents a method for inferring spatial gene expression from hematoxylin and eosin stains. The proposed method offers a data-driven approach to analyzing histopathology images without relying on expert annotations and could be a valuable tool for cancer screening and diagnosis in the clinics. Paper II introduces a method for jointly modeling spatial gene expression with histology images. We show that the method can predict super-resolved gene expression and transcriptionally characterize small-scale anatomical structures. Paper III proposes a method for learning flexible Markov kernels to model continuous and discrete data distributions. We demonstrate the method on various image synthesis tasks, including unconditional image generation and inpainting. Paper IV leverages the techniques introduced in Paper III to integrate data from different spatial biology experiments. The proposed method can be used for data imputation, super resolution, and cross-modality data transfer.
|
|
| 7. |
- Frederiksen, Line Elmerdahl, et al.
(författare)
-
Psychiatric disorders in childhood cancer survivors in Denmark, Finland, and Sweden : a register-based cohort study from the SALiCCS research programme
- 2022
-
Ingår i: The Lancet Psychiatry. - 2215-0366 .- 2215-0374. ; 69
-
Tidskriftsartikel (refereegranskat)abstract
- Background: A childhood cancer diagnosis and treatment-induced somatic late effects can affect the long-term mental health of survivors. We aimed to explore whether childhood cancer survivors are at higher risk of psychiatric disorders later in life than their siblings and the general population. Methods: In this register-based cohort study (part of the Socioeconomic Consequences in Adult Life after Childhood Cancer [SALiCCS] research programme), we included 5-year survivors of childhood cancer diagnosed before 20 years of age between Jan 1, 1974 and Dec 31, 2011, in Denmark, Finland, and Sweden. In Denmark and Sweden, 94·7% of individuals were born in a Nordic country (ie, Denmark, Finland, Iceland, Norway, or Sweden); similar information was not available in Finland. Data on ethnicity were not collected. Survivors were compared with their siblings and randomly selected individuals from the general population who were matched to the survivors by year of birth, sex, and geographical region. We followed up our study population from 5 years after the childhood cancer diagnosis or corresponding calendar date for matched individuals (the index date) until Aug 11, 2017, and assessed information on hospital contacts for any and specific psychiatric disorders. For siblings, the index date was defined as 5 years from the date on which they were of the same age as their sibling survivor when diagnosed with cancer. Findings: The study population included 18 621 childhood cancer survivors (9934 [53·3%] males and 8687 [46·7%] females), 24 775 siblings (12 594 [50·8%] males and 12 181 [49·2%] females), and 88 630 matched individuals (47 300 [53·4%] males and 41 330 [46·6%] females). The cumulative incidence proportion of having had a psychiatric hospital contact by 30 years of age between Jan 1, 1979, and Aug 11, 2017, was 15·9% (95% CI 15·3–16·5) for childhood cancer survivors, 14·0% (13·5–14·5) for siblings, and 12·7% (12·4–12·9) for matched individuals. Despite a small absolute difference, survivors were at higher relative risk of any psychiatric hospital contact than their siblings (1·39, 1·31–1·48) and matched individuals (hazard ratio 1·34, 95% CI 1·28–1·39). The higher risk persisted at the age of 50 years. Survivors had a higher burden of recurrent psychiatric hospital contacts and had more hospital contacts for different psychiatric disorders than their siblings and the matched individuals. Interpretation: Childhood cancer survivors are at higher long-term risk of psychiatric disorders than their siblings and matched individuals from the general population. To improve mental health and the overall quality of life after childhood cancer, survivorship care should include a focus on early signs of mental health problems, especially among high-risk groups of survivors. Funding: NordForsk, Aarhus University, Swedish Childhood Cancer Foundation, Danish Health Foundation, and Swiss National Science Foundation.
|
|
| 8. |
- Ott, Simon, et al.
(författare)
-
ThoughtSource : A central hub for large language model reasoning data
- 2023
-
Ingår i: Scientific Data. - : Springer Nature. - 2052-4463. ; 10:1
-
Tidskriftsartikel (refereegranskat)abstract
- Large language models (LLMs) such as GPT-4 have recently demonstrated impressive results across a wide range of tasks. LLMs are still limited, however, in that they frequently fail at complex reasoning, their reasoning processes are opaque, they are prone to 'hallucinate' facts, and there are concerns about their underlying biases. Letting models verbalize reasoning steps as natural language, a technique known as chain-of-thought prompting, has recently been proposed as a way to address some of these issues. Here we present ThoughtSource, a meta-dataset and software library for chain-of-thought (CoT) reasoning. The goal of ThoughtSource is to improve future artificial intelligence systems by facilitating qualitative understanding of CoTs, enabling empirical evaluations, and providing training data. This first release of ThoughtSource integrates seven scientific/medical, three general-domain and five math word question answering datasets.
|
|
| 9. |
- Rapin, Nicolas, et al.
(författare)
-
Comparing cancer vs normal gene expression profiles identifies new disease entities and common transcriptional programs in AML patients
- 2014
-
Ingår i: Blood. - : American Society of Hematology. - 1528-0020 .- 0006-4971. ; 123:6, s. 894-904
-
Tidskriftsartikel (refereegranskat)abstract
- Gene expression profiling has been used extensively to characterize cancer, identify novel subtypes, and improve patient stratification. However, it has largely failed to identify transcriptional programs that differ between cancer and corresponding normal cells and has not been efficient in identifying expression changes fundamental to disease etiology. Here we present a method that facilitates the comparison of any cancer sample to its nearest normal cellular counterpart, using acute myeloid leukemia (AML) as a model. We first generated a gene expression-based landscape of the normal hematopoietic hierarchy, using expression profiles from normal stem/progenitor cells, and next mapped the AML patient samples to this landscape. This allowed us to identify the closest normal counterpart of individual AML samples and determine gene expression changes between cancer and normal. We find the cancer vs normal method (CvN method) to be superior to conventional methods in stratifying AML patients with aberrant karyotype and in identifying common aberrant transcriptional programs with potential importance for AML etiology. Moreover, the CvN method uncovered a novel poor-outcome subtype of normal-karyotype AML, which allowed for the generation of a highly prognostic survival signature. Collectively, our CvN method holds great potential as a tool for the analysis of gene expression profiles of cancer patients.
|
|
| 10. |
|
|
