| 1. |
|
|
| 2. |
|
|
| 3. |
|
|
| 4. |
|
|
| 5. |
|
|
| 6. |
|
|
| 7. |
- Romagnoni, A, et al.
(författare)
-
Comparative performances of machine learning methods for classifying Crohn Disease patients using genome-wide genotyping data
- 2019
-
Ingår i: Scientific reports. - : Springer Science and Business Media LLC. - 2045-2322. ; 9:1, s. 10351-
-
Tidskriftsartikel (refereegranskat)abstract
- Crohn Disease (CD) is a complex genetic disorder for which more than 140 genes have been identified using genome wide association studies (GWAS). However, the genetic architecture of the trait remains largely unknown. The recent development of machine learning (ML) approaches incited us to apply them to classify healthy and diseased people according to their genomic information. The Immunochip dataset containing 18,227 CD patients and 34,050 healthy controls enrolled and genotyped by the international Inflammatory Bowel Disease genetic consortium (IIBDGC) has been re-analyzed using a set of ML methods: penalized logistic regression (LR), gradient boosted trees (GBT) and artificial neural networks (NN). The main score used to compare the methods was the Area Under the ROC Curve (AUC) statistics. The impact of quality control (QC), imputing and coding methods on LR results showed that QC methods and imputation of missing genotypes may artificially increase the scores. At the opposite, neither the patient/control ratio nor marker preselection or coding strategies significantly affected the results. LR methods, including Lasso, Ridge and ElasticNet provided similar results with a maximum AUC of 0.80. GBT methods like XGBoost, LightGBM and CatBoost, together with dense NN with one or more hidden layers, provided similar AUC values, suggesting limited epistatic effects in the genetic architecture of the trait. ML methods detected near all the genetic variants previously identified by GWAS among the best predictors plus additional predictors with lower effects. The robustness and complementarity of the different methods are also studied. Compared to LR, non-linear models such as GBT or NN may provide robust complementary approaches to identify and classify genetic markers.
|
|
| 8. |
- Abdalla, H., et al.
(författare)
-
First limits on the very-high energy gamma-ray afterglow emission of a fast radio burst HESS observations of FRB 150418
- 2017
-
Ingår i: Astronomy and Astrophysics. - : The European Southern Observatory (ESO). - 0004-6361 .- 1432-0746. ; 597
-
Tidskriftsartikel (refereegranskat)abstract
- Aims. Following the detection of the fast radio burst FRB150418 by the SUPERB project at the Parkes radio telescope, we aim to search for very-high energy gamma-ray afterglow emission. Methods. Follow-up observations in the very-high energy gamma-ray domain were obtained with the H.E.S.S. imaging atmospheric Cherenkov telescope system within 14.5 h of the radio burst. Results. The obtained 1.4 h of gamma-ray observations are presented and discussed. At the 99% C.L. we obtained an integral upper limit on the gamma-ray flux of Phi(gamma)(E > 350 GeV) < 1.33 x 10(-8) m(-2) s(-1). Differential flux upper limits as function of the photon energy were derived and used to constrain the intrinsic high-energy afterglow emission of FRB 150418. Conclusions. No hints for high-energy afterglow emission of FRB 150418 were found. Taking absorption on the extragalactic background light into account and assuming a distance of z = 0 : 492 based on radio and optical counterpart studies and consistent with the FRB dispersion, we constrain the gamma-ray luminosity at 1 TeV to L < 5 : 1 x 10(47) erg/s at 99% C.L.
|
|
| 9. |
|
|
| 10. |
- Aguiar, A., et al.
(författare)
-
Practices in prescribing protein substitutes for PKU in Europe : No uniformity of approach
- 2015
-
Ingår i: Molecular Genetics and Metabolism. - : Elsevier BV. - 1096-7192 .- 1096-7206. ; 115:1, s. 17-22
-
Tidskriftsartikel (refereegranskat)abstract
- Background: There appears little consensus concerning protein requirements in phenylketonuria (PKU). Methods: A questionnaire completed by 63 European and Turkish IMD centres from 18 countries collected data on prescribed total protein intake (natural/intact protein and phenylalanine-free protein substitute [PS]) by age, administration frequency and method, monitoring, and type of protein substitute. Data were analysed by European region using descriptive statistics. Results: The amount of total protein (from PS and natural/intact protein) varied according to the European region. Higher median amounts of total protein were prescribed in infants and children in Northern Europe (n = 24 centres) (infants <1 year, >2-3 g/kg/day; 1-3 years of age, >2-3 g/kg/day; 4-10 years of age, >1.5-2.5 g/kg/day) and Southern Europe (n = 10 centres) (infants <1 year, 2.5 g/kg/day, 1-3 years of age, 2 g/kg/day; 4-10 years of age, 1.5-2 g/kg/day), than by Eastern Europe (n = 4 centres) (infants <1 year, 2.5 g/kg/day, 1-3 years of age, >2-2.5 g/kg/day; 4-10 years of age, >1.5-2 g/kg/day) and with Western Europe (n = 25 centres) giving the least (infants <1 year, >2-2.5 g/kg/day, 1-3 years of age, 1.5-2 g/kg/day; 4-10 years of age, 1-1.5 g/kg/day). Total protein prescription was similar in patients aged >10 years (1-1.5 g/kg/day) and maternal patients (1-1.5 g/kg/day). Conclusions: The amounts of total protein prescribed varied between European countries and appeared to be influenced by geographical region. In PKU, all gave higher than the recommended 2007 WHO/FAO/UNU safe levels of protein intake for the general population.
|
|
