Search

Vertebral labelling and segmentation are two fundamental tasks in an automated spine processing pipeline. Reliable and accurate processing of spine images is expected to benefit clinical decision-support systems for diagnosis, surgery planning, and population-based analysis on spine and bone health. However, designing automated algorithms for spine processing is challenging predominantly due to considerable variations in anatomy and acquisition protocols and due to a severe shortage of publicly available data. Addressing these limitations, the Large Scale Vertebrae Segmentation Challenge (VerSe) was organised in conjunction with the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) in 2019 and 2020, with a call for algorithms towards labelling and segmentation of vertebrae. Two datasets containing a total of 374 multi-detector CT scans from 355 patients were prepared and 4505 vertebrae have individually been annotated at voxel-level by a human-machine hybrid algorithm (https://osf.io/nqjyw/, https://osf.io/t98fz/). A total of 25 algorithms were benchmarked on these datasets. In this work, we present the the results of this evaluation and further investigate the performance-variation at vertebra-level, scan-level, and at different fields-of-view. We also evaluate the generalisability of the approaches to an implicit domain shift in data by evaluating the top performing algorithms of one challenge iteration on data from the other iteration. The principal takeaway from VerSe: the performance of an algorithm in labelling and segmenting a spine scan hinges on its ability to correctly identify vertebrae in cases of rare anatomical variations. The content and code concerning VerSe can be accessed at: https://github.com/anjany/verse.

International audience

International audience; A rapid method for the identification of epistatic ‘dormant’ SNPs We present a unique computational approach for the identification of epistatic SNPs based on SNPs with significant yet opposed effects depending on the genetic background. We introduce the mechanical heuristics of the approach based on first, binning the population according to their genomic-estimated breeding value (GEBV) and second, performing genome-wide association studies (GWAS) within each bin. SNPs are deemed to be epistatic if significant but with different signed effects in the GWAS from the most extreme bins containing individuals with the lowest and highest GEBV. We then show that these heuristics are equivalent to a regression of residuals on GEBV. Next, we illustrate our approach with a dataset of 2,111 cattle genotyped for 651,253 SNPs and using yearling weight as the phenotype. We identify 243 epistatic SNPs, and argue that these SNPs are ‘dormant’ with an additive effect waiting to be ‘released’ if selection moves the population to either tail of the genetic value distribution.

Einleitung und Fragestellung: Die Versorgung von Patienten mit psychischen Erkrankungen im höheren Lebensalter, insbesondere Demenzformen, wird in Zukunft in Thüringen verstärkt zu einem erheblichen Problem [ref:1]. Viele Patienten leben in ländlichen Regionen Thüringens[for full text, please go to the a.m. URL], GMDS 2014; 59. Jahrestagung der Deutschen Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie e.V. (GMDS)

High intramuscular fat (IMF) awards price premiums to beef producers and is associated with meat quality and flavor. Studying gene interactions and pathways that affect IMF might unveil causative physiological mechanisms and inform genomic selection, leading to increased accuracy of predictions of breeding value. To study gene interactions and pathways, a gene network was derived from genetic markers associated with direct measures of IMF, other fat phenotypes, feedlot performance, and a number of meat quality traits relating to body conformation, development, and metabolism that might be plausibly expected to interact with IMF biology. Marker associations were inferred from genomewide association studies (GWAS) based on high density genotypes and 29 traits measured on 10,181 beef cattle animals from 3 breed types. For the network inference, SNP pairs were assessed according to the strength of the correlation between their additive association effects across the 29 traits. The co-association inferred network was formed by 2,434 genes connected by 28,283 edges. Topological network parameters suggested a highly cohesive network, in which the genes are strongly functionally interconnected. Pathway and network analyses pointed towards a trio of transcription factors (TF) as key regulators of carcass IMF: PPARG-C1A, HNF4G, and FOXP3. Importantly, none of these genes would have been deemed as significantly associated with IMF from the GWAS. Instead, a total of 313 network genes show significant co-association with the 3 TF. These genes belong to a wide variety of biological functions, canonical pathways, and genetic networks linked to IMF-related phenotypes. In summary, our GWAS and network predictions are supported by the current literature and suggest a cooperative role for the 3 TF and other interacting genes including CAPN6, STC2, MAP2K4, EYA1, COPS5, XKR4, NR2E1, TOX, ATF1, ASPH, TGS1, and TTPA as modulators of carcass and meat quality traits in beef cattle.

Despite major advances in deciphering the neuropathological hallmarks of amyotrophic lateral sclerosis (ALS), validated neurochemical biomarkers for monitoring disease activity, earlier diagnosis, defining prognosis and unlocking key pathophysiological pathways are lacking. Although several candidate biomarkers exist, translation into clinical application is hindered by small sample numbers, especially longitudinal, for independent verification. This review considers the potential routes to the discovery of neurochemical markers in ALS, and provides a consensus statement on standard operating procedures that will facilitate multicenter collaboration, validation and ultimately clinical translation.