Closing the translation gap: AI applications in digital pathology

Recent advances in artificial intelligence show tremendous promise to improve the accuracy, reproducibility, and availability of medical diagnostics across a number of medical subspecialities. This is especially true in the field of digital pathology, which has recently witnessed a surge in publications describing state-of-the-art performance for machine learning models across a wide range of diagnostic applications. Nonetheless, despite this promise, there remain significant gaps in translating applications for any of these technologies into actual clinical practice. In this review, we will first give a brief overview of the recent progress in applying AI to digitized pathology images, focusing on how these tools might be applied in clinical workflows in the near term to improve the accuracy and efficiency of pathologists. Then we define and describe in detail the various factors that need to be addressed in order to successfully close the “translation gap” for AI applications in digital pathology.     

The Genomic Architecture of Adaptation to Larval Malnutrition Points to a Trade-off with Adult Starvation Resistance in Drosophila

Periods of nutrient shortage impose strong selection on animal populations. Experimental studies of genetic adaptation to nutrient shortage largely focus on resistance to acute starvation at adult stage; it is not clear how conclusions drawn from these studies extrapolate to other forms of nutritional stress. We studied the genomic signature of adaptation to chronic juvenile malnutrition in six populations of Drosophila melanogaster evolved for 150 generations on an extremely nutrient-poor larval diet. Comparison with control populations evolved on standard food revealed repeatable genomic differentiation between the two set of population, involving >3,000 candidate SNPs forming >100 independently evolving clusters. The candidate genomic regions were enriched in genes implicated in hormone, carbohydrate, and lipid metabolism, including some with known effects on fitness-related life-history traits. Rather than being close to fixation, a substantial fraction of candidate SNPs segregated at intermediate allele frequencies in all malnutrition-adapted populations. This, together with patterns of among-population variation in allele frequencies and estimates of Tajima’s D, suggests that the poor diet results in balancing selection on some genomic regions. Our candidate genes for tolerance to larval malnutrition showed a high overlap with genes previously implicated in acute starvation resistance. However, adaptation to larval malnutrition in our study was associated with reduced tolerance to acute adult starvation. Thus, rather than reflecting synergy, the shared genomic architecture appears to mediate an evolutionary trade-off between tolerances to these two forms of nutritional stress.     

Correction to: Transmission of a novel predatory behaviour is not restricted to kin

Biological Invasions - A correction to this paper has been published: https://doi.org/10.1007/s10530-021-02548-x     

Advances in approaches to seagrass restoration in Australia

Three case studies involving two temperate Australian seagrass species – Pondweed (Ruppia tuberosa) and Ribbon Weed (Posidonia australis) – highlight different approaches to their restoration. Seeds and rhizomes were used in three collaborative programmes to promote new approaches to scale up restoration outcomes.     

An assessment of the seascape genetic structure and hydrodynamic connectivity for subtropical seagrass restoration

Seagrass ecosystems have suffered significant declines globally and focus is shifting to restoration efforts. A key component to successful restoration is an understanding of the genetic factors potentially influencing restoration success. This includes understanding levels of connectivity between restoration locations and neighboring seagrass populations that promote natural recovery (source and sink populations), the identification of potential donor populations, and assessment of genetic diversity of restored meadows and material used for restoration. In this study, we carry out genetic surveys of 352 individuals from 13 populations using 11 polymorphic microsatellite loci to inform seagrass restoration activities by: (1) understanding levels of genetic and genotypic diversity within meadows; and (2) understanding genetic structure and patterns of connectivity among these meadows to determine which source sites may be most appropriate to assist recovery of three restoration sites. The study identified high genotypic diversity within the locations analyzed from the Port of Gladstone and Rodd's Bay region, indicating sexual reproduction is important in maintaining populations. Overall, we detected significant genetic structuring among sites with the Bayesian structure analysis identifying genetic clusters that largely conformed to a northern, central, and southern region. This suggests limited gene flow between regions, although there does appear to be some connectivity within regions. The hydrodynamic models showed that seeds were largely locally retained, while fragments were more widely dispersed. Limited gene flow between regions suggests donor material for restoration should be sourced locally where possible.     

Examining the reliability of the flexor carpi radialis V-wave at different levels of muscle contraction

This study examined the reliability and scaling of the flexor carpi radialis (FCR) V-wave during submaximal and maximal voluntary muscle contractions (MVC). 23 participants were tested on three separate sessions. For each session, participants performed isometric wrist flexions at five contraction levels (20, 40, 60, 80 and 100 %MVC). When the target contraction level was reached, a supramaximal electrical stimulus was applied to the median nerve in order to elicit an FCR V-wave. Across all participants, the FCR V-wave amplitude, normalized to its superimposed M-wave amplitude, increased from 0.030 ± 0.001 to 0.143 ± 0.015 (P < 0.001) as the muscle contraction increased from 20 to 100 %MVC. Contraction level did not influence the reliability of evoking the FCR V-wave, as the V-wave demonstrated both stability and consistency. With the exception of a single day main effect during the 20 %MVC condition, V:M_sup was not different across days or trials (P > 0.05) indicating measurement stability. High reliability co-efficients (0.827–0.913) at each contraction level signified measurement consistency. This study establishes that FCR V-waves can be reliably evoked during both submaximal and maximal muscle contractions and suggests the possibility for FCR V-wave recordings to be used to document neuromuscular adaptations associated with factors such as training or fatigue.     

The effects of age,energy and protein intake on protein turnover and the expression of proteolysis-related genes in the broiler breeder hen

A study was conducted to determine the changes that occur to proteolysis and related genes due to age, protein, and energy intake in high-yield broiler breeder hens (Gallus gallus). Cobb 700 broiler breeders were randomly assigned to one of six diets in a 2 × 3 factorial fashion. Two levels of energy (390 and 450 kcal/day) and three levels of protein (22, 24, and 26 g CP/day) were utilized. Protein turnover was determined in the left pectoralis at 22, 26, 31 and 44 weeks. Relative mRNA expression of calpain 2 (CAPN2), proteasome C2 subunit (PSMA1), and F box protein 32 (FBXO32) were determined via RT-PCR at 20, 25, and 44 weeks. Contrasts indicate fractional synthesis rate (FSR) and FBXO32 increase to a maximum at 25–26 weeks and a decrease thereafter. A significant drop in PSMA1 and FBXO32 was observed between 25 and 44 weeks and matched the decrease observed in FBR. No differences were detected in the levels of fractional synthesis and degradation, or the expression of CAPN2, PSMA1, and FBXO32, due to protein or energy intake. In summary, protein turnover was upregulated during the transition into sexual maturity and decreased thereafter. The observed changes in degradation appeared to be mediated by the ubiquitin–proteasome pathway.