Genomic prediction using machine learning: a comparison of the performance of regularized regression, ensemble, instance-based and deep learning methods on synthetic and empirical data.

BACKGROUND: The accurate prediction of genomic breeding values is central to genomic selection in both plant and animal breeding studies. Genomic prediction involves the use of thousands of molecular markers spanning the entire genome and therefore requires methods able to efficiently handle high dimensional data. Not surprisingly, machine learning methods are becoming widely advocated for and used in genomic prediction studies. These methods encompass different groups of supervised and unsupervised learning methods. Although several studies have compared the predictive performances of individual methods, studies comparing the predictive performance of different groups of methods are rare. However, such studies are crucial for identifying (i) groups of methods with superior genomic predictive performance and assessing (ii) the merits and demerits of such groups of methods relative to each other and to the established classical methods. Here, we comparatively evaluate the genomic predictive performance and informally assess the computational cost of several groups of supervised machine learning methods, specifically, regularized regression methods, deep, ensemble and instance-based learning algorithms, using one simulated animal breeding dataset and three empirical maize breeding datasets obtained from a commercial breeding program. RESULTS: Our results show that the relative predictive performance and computational expense of the groups of machine learning methods depend upon both the data and target traits and that for classical regularized methods, increasing model complexity can incur huge computational costs but does not necessarily always improve predictive accuracy. Thus, despite their greater complexity and computational burden, neither the adaptive nor the group regularized methods clearly improved upon the results of their simple regularized counterparts. This rules out selection of one procedure among machine learning methods for routine use in genomic prediction. The results also show that, because of their competitive predictive performance, computational efficiency, simplicity and therefore relatively few tuning parameters, the classical linear mixed model and regularized regression methods are likely to remain strong contenders for genomic prediction. CONCLUSIONS: The dependence of predictive performance and computational burden on target datasets and traits call for increasing investments in enhancing the computational efficiency of machine learning algorithms and computing resources.

Repeatability of whole herd lameness scoring: an analysis of a New Zealand dataset.

AIMS: To assess whether a whole-herd lameness score on a New Zealand dairy farm in spring could predict lameness prevalence on the same farm in summer (and vice versa) and whether a single-herd lameness score could be used to determine whether herd lameness prevalence was < 5% in both spring and summer. METHODS: Prevalence data (proportion of the herd with lameness score ≥ 2 and with score 3; 0-3 scale) from a study where 120 dairy farms across New Zealand were scored in spring and in the following summer were analysed using limits-of-agreement analysis. In addition, farms were categorised as having either acceptable welfare (lameness prevalence < 5% in both spring and summer) or not (lameness prevalence ≥ 5% in either spring or summer or both). The accuracy and specificity of a single, whole-herd lameness score at identifying herds with acceptable welfare were then calculated. RESULTS: The limits-of-agreement analysis suggests that 95% of the time, the prevalence of lameness in summer would be expected to be between 0.23 and 4.3 times that of the prevalence in spring. The specificity and accuracy of identifying a farm as acceptable on both occasions from a single observation were, respectively, 74% and 92% in spring, and 59% and 87% in summer. CONCLUSIONS: A single, one-off, whole-herd lameness score does not accurately predict future lameness prevalence. Similarly, acceptable status (lameness prevalence < 5%) in one season is not sufficiently specific to be used to predict welfare status in subsequent seasons. CLINICAL RELEVANCE: Whole-herd lameness scoring should be used principally as a means of detecting lame cows for treatment. A single whole-herd lameness score by an independent assessor should not be used to determine a herd's welfare status.

Differential utility of various frailty diagnostic tools in non-geriatric hospital departments of several countries: A longitudinal study.

BACKGROUND: There is limited knowledge on the performance of different frailty scales in clinical settings. We sought to evaluate in non-geriatric hospital departments the feasibility, agreement and predictive ability for adverse events after 1 year follow-up of several frailty assessment tools. METHODS: Longitudinal study with 667 older adults recruited from five hospitals in three different countries (Spain, Italy and United Kingdom). Participants were older than 75 years attending the emergency room, cardiology and surgery departments. Frailty scales used were Frailty Phenotype (FP), FRAIL scale, Tilburg and Groningen Frailty Indicators, and Clinical Frailty Scale (CFS). Analyses included the prevalence of frailty, degree of agreement between tools, feasibility and prognostic value for hospital readmission, worsening of disability and mortality, by tool and setting. RESULTS: Emergency Room and cardiology were the settings with the highest frailty prevalence, varying by tool between 40.4% and 67.2%; elective surgery was the one with the lowest prevalence (between 13.2% and 38.2%). The tools showed a fair to moderate agreement. FP showed the lowest feasibility, especially in urgent surgery (35.6%). FRAIL, CFS and FP predicted mortality and readmissions in several settings, but disability worsening only in cardiology. CONCLUSIONS: Frailty is a highly frequent condition in older people attending non-geriatric hospital departments. We recommend that based upon their current feasibility and predictive ability, the FRAIL scale, CFS and FP should be preferentially used in these settings. The low concordance among the tools and differences in prevalence reported and predictive ability suggest the existence of different subtypes of frailty.

Integrating genome-wide association study into genomic selection for the prediction of agronomic traits in rice (Oryza sativa L.).

Accurately identifying varieties with targeted agronomic traits was thought to contribute to genetic selection and accelerate rice breeding progress. Genomic selection (GS) is a promising technique that uses markers covering the whole genome to predict the genomic-estimated breeding values (GEBV), with the ability to select before phenotypes are measured. To choose the appropriate GS models for breeding work, we analyzed the predictability of nine agronomic traits measured from a population of 459 diverse rice varieties. By the comparison of eight representative GS models, we found that the prediction accuracies ranged from 0.407 to 0.896, with reproducing kernel Hilbert space (RKHS) having the highest predictive ability in most traits. Further results demonstrated the predictivity of GS is altered by several factors. Moreover, we assessed the method of integrating genome-wide association study (GWAS) into various GS models. The predictabilities of GS combined peak-associated markers generated from six different GWAS models were significantly different; a recommendation of Mixed Linear Model (MLM)-RKHS was given for the GWAS-GS-integrated prediction. Finally, based on the above result, we experimented with applying the P-values obtained from optimal GWAS models into ridge regression best linear unbiased prediction (rrBLUP), which benefited the low predictive traits in rice. Supplementary Information: The online version contains supplementary material available at 10.1007/s11032-023-01423-y.

Predictive ability of traditional and novel anthropometric measurement indices for cardio-metabolic diseases in Chinese adults: China Health and Nutrition Survey (CHNS) cohort study.

BACKGROUND AND AIMS: Cardio-metabolic diseases has been shown to be strongly associated with obesity. The aim of this study was to compare the predictive value of traditional and novel anthropometric measurement indices for cardio-metabolic diseases risk and evaluate whether new indicators can provide important information in addition to traditional indicators. METHODS AND RESULTS: China Health and Nutrition Survey (CHNS) data were obtained for this study. Baseline information for healthy participants was gathered from 1997 to 2004. The incidence of cardio-metabolic diseases was collected from 2009 to 2015 for cohort analysis. The predictive ability of each index for the risk of cardio-metabolic diseases was evaluated with time-dependent ROC analysis. Body mass index (BMI) showed the greatest predictive ability for cardio-metabolic disease incidence among all traditional and novel indices (Harrell's C statistic (95% CI): 0.7386 (0.7266-0.7507) for hypertension, 0.7496 (0.7285-0.7706) for diabetes, 0.7895 (0.7593-0.8196) for stroke and 0.7581 (0.7193-0.7969) for myocardial infarction). The addition of novel indices separately into the BMI model did not improve the predictive ability. Novel anthropometric measurement indices such as a body shape index (ABSI), abdominal volume index (AVI) and triponderal mass index (TMI), had a certain prediction ability for adults with BMI <24 kg/m2 compared to those with BMI ≥24 kg/m2. CONCLUSION: No strong evidence supports novel anthropometric measurement indices were better than BMI in the prediction of cardio-metabolic diseases incidence among Chinese adults. Novel anthropometric measurement indices, mainly for abdominal obesity, may have a high predictive effect for adults with BMI <24 kg/m2.

The Predictive Ability of C-Peptide in Distinguishing Type 1 Diabetes From Type 2 Diabetes: A Systematic Review and Meta-Analysis.

OBJECTIVE: This systematic review and meta-analysis aimed to investigate the predictive ability of plasma connecting peptide (C-peptide) levels in discriminating type 1 diabetes (T1D) from type 2 diabetes (T2D) and to inform evidence-based guidelines in diabetes classification. METHODS: We conducted a holistic review and meta-analysis using PubMed, MEDLINE, EMBASE, and Scopus. The citations were screened from 1942 to 2021. The quality criteria and the preferred reporting items for systematic reviews and meta-analysis checklist were applied. The protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42022355088). RESULTS: A total of 23,658 abstracts were screened and 46 full texts reviewed. Of the 46 articles screened, 12 articles were included for the meta-analysis. Included studies varied by race, age, time, and proportion of individuals. The main outcome measure in all studies was C-peptide levels. A significant association was reported between C-peptide levels and the classification and diagnosis of diabetes. Furthermore, lower concentrations and the cutoff of <0.20 nmol/L for fasting or random plasma C-peptide was indicative of T1D. In addition, this meta-analysis revealed the predictive ability of C-peptide levels in discriminating T1D from T2D. Results were consistent using both fixed- and random-effect models. The I2 value (98.8%) affirmed the variability in effect estimates was due to heterogeneity rather than sampling error among all selected studies. CONCLUSION: Plasma C-peptide levels are highly associated and predictive of the accurate classification and diagnosis of diabetes types. A plasma C-peptide cutoff of ≤0.20 mmol/L is indicative of T1D and of ≥0.30 mmol/L in the fasting or random state is indicative of T2D.

Effectiveness of mid-infrared spectroscopy for the prediction of cow milk metabolites.

Proton nuclear magnetic resonance (1H NMR) spectroscopy is acknowledged as one of the most powerful analytical methods with cross-cutting applications in dairy foods. To date, the use of 1H NMR spectroscopy for the collection of milk metabolic profile is hindered by costly and time-consuming sample preparation and analysis. The present study aimed at evaluating the accuracy of mid-infrared spectroscopy (MIRS) as a rapid method for the prediction of cow milk metabolites determined through 1H NMR spectroscopy. Bulk milk (n = 72) and individual milk samples (n = 482) were analyzed through one-dimensional 1H NMR spectroscopy and MIRS. Nuclear magnetic resonance spectroscopy identified 35 milk metabolites, which were quantified in terms of relative abundance, and MIRS prediction models were developed on the same 35 milk metabolites, using partial least squares regression analysis. The best MIRS prediction models were developed for galactose-1-phosphate, glycerophosphocholine, orotate, choline, galactose, lecithin, glutamate, and lactose, with coefficient of determination in external validation from 0.58 to 0.85, and ratio of performance to deviation in external validation from 1.50 to 2.64. The remaining 27 metabolites were poorly predicted. This study represents a first attempt to predict milk metabolome. Further research is needed to specifically address whether developed prediction models may find practical application in the dairy sector, with particular regard to the screening of dairy cows' metabolic status, the quality control of dairy foods, and the identification of processed milk or incorrectly stored milk.

Heritability estimates and predictive ability for pig meat quality traits using identity-by-state and identity-by-descent relationships in an F2 population.

Genomic relationships can be computed with dense genome-wide genotypes through different methods, either based on identity-by-state (IBS) or identity-by-descent (IBD). The latter has been shown to increase the accuracy of both estimated relationships and predicted breeding values. However, it is not clear whether an IBD approach would achieve greater heritability ( h 2 ) and predictive ability ( r ̂ y , y ̂ ) than its IBS counterpart for data with low-depth pedigrees. Here, we compare both approaches in terms of the estimated of h 2 and r ̂ y , y ̂ , using data on meat quality and carcass traits recorded in experimental crossbred pigs, with a pedigree constrained to only three generations. Three animal models were fitted which differed on the relationship matrix: an IBS model ( G IBS ), an IBD (defined within the known pedigree) model ( G IBD ), and a pedigree model ( A 22 ). In 9 of 20 traits, the range of increase for the estimates of σ u 2 and h 2 was 1.2-2.9 times greater with G IBS and G IBD models than with A 22 . Whereas for all traits, both parameters were similar between genomic models. The r ̂ y , y ̂ of the genomic models was higher compared to A 22 . A scarce increment in r ̂ y , y ̂ was found with G IBS when compared to G IBD , most likely due to the former recovering sizeable relationships among founder F0 animals.

Neutrophil Percentage to Albumin Ratio was Associated with Clinical Outcomes in Coronary Care Unit Patients.

Background: Neutrophil percentage to albumin ratio (NPAR) has been shown to be correlated with the prognosis of various diseases. This study aimed to explore the effect of NPAR on the prognosis of patients in coronary care units (CCU). Method: All data in this study were extracted from the Medical Information Mart for Intensive Care III (MIMIC-III, version1.4) database. All patients were divided into four groups according to their NPAR quartiles. The primary outcome was in-hospital mortality. Secondary outcomes were 30-day mortality, 365-day mortality, length of CCU stay, length of hospital stay, acute kidney injury (AKI), and continuous renal replacement therapy (CRRT). A multivariate binary logistic regression analysis was performed to confirm the independent effects of NPAR. Cox regression analysis was performed to analyze the association between NPAR and 365-day mortality. The curve in line with overall trend was drawn by local weighted regression (Lowess). Subgroup analysis was used to determine the effect of NPAR on in-hospital mortality in different subgroups. Receiver operating characteristic (ROC) curves were used to evaluate the ability of NPAR to predict in-hospital mortality. Kaplan-Meier curves were constructed to compare the cumulative survival rates among different groups. Result: A total of 2364 patients in CCU were enrolled in this study. The in-hospital mortality rate increased significantly as the NPAR quartiles increased (p < 0.001). In multivariate logistic regression analysis, NPAR was independently associated with in-hospital mortality (quartile 4 versus quartile 1: odds ratio [OR], 95% confidence interval [CI]: 1.83, 1.20-2.79, p = 0.005, p for trend < 0.001). In Cox regression analysis, NPAR was independently associated with 365-day mortality (quartile 4 versus quartile 1: OR, 95% CI: 1.62, 1.16-2.28, p = 0.005, p for trend < 0.001). The Lowess curves showed a positive relationship between NPAR and in-hospital mortality. The moderate ability of NPAR to predict in-hospital mortality was demonstrated through ROC curves. The area under the curves (AUC) of NPAR was 0.653 (p < 0.001), which is better than that of the platelet to lymphocyte ratio (PLR) (p < 0.001) and neutrophil count (p < 0.001) but lower than the Sequential Organ Failure Assessment (p = 0.046) and Simplified Acute Physiology Score II (p < 0.001). Subgroup analysis did not reveal any obvious interactions in most subgroups. However, Kaplan-Meier curves showed that as NPAR quartiles increased, the 30-day (log-rank, p < 0.001) and 365-day (log-rank, p < 0.001) cumulative survival rates decreased significantly. NPAR was also independently associated with AKI (quartile 4 versus quartile 1: OR, 95% CI: 1.57, 1.19-2.07, p = 0.002, p for trend = 0.001). The CCU and hospital stay length was significantly prolonged in the higher NPAR quartiles. Conclusions: NPAR is an independent risk factor for in-hospital mortality in patients in CCU and has a moderate ability to predict in-hospital mortality.

Assessment of the predictive ability of the Johns Hopkins Fall Risk Assessment Tool (Chinese Version) in inpatient settings.

AIMS: This study was to assess the predictive ability of the Johns Hopkins Fall Risk Assessment Tool (Chinese Version) in inpatient settings. DESIGN: A case-control study. METHODS: This study was conducted in a tertiary hospital based on 2019 data. With a case-control design in a 1:2 ratio, the predictive ability of the Johns Hopkins Fall Risk Assessment Tool (Chinese Version) was determined by ROC curve. The best cut point was identified based on sensitivity, specificity, positive predict value and negative predict value. Conditional logistical regression analysis was conducted to test the predictive ability of each indicator. RESULTS: The study included 309 patients, with 103 in the case group and 206 in the control groups. Generally, the predictive ability was acceptable with the area under ROC curve value at 0.73 (95% CI: 0.67-0.79). Positive predict value and negative predict value performed best at the cut point of 13. Sensitivity at cut point 6 was much higher than that at cut point 13, though specificity was lower. Except for age, all indicators in the Johns Hopkins Fall Risk Assessment Tool (Chinese Version) demonstrated significant predictive ability as to occurrence of fall. CONCLUSION: The Johns Hopkins Fall Risk Assessment Tool (Chinese Version) is a reliable assessment instrument in the inpatient settings. IMPACT: This is the first study that evaluated the predictive ability of the Johns Hopkins Fall Risk Assessment Tool (Chinese version) in the inpatient settings, and proved that the instrument is reliable for assessing inpatient fall risks. Further studies could be carried out to assess the predict ability of Johns Hopkins Fall Risk Assessment Tool (Chinese version) among specific populations.

Genetic evaluation of age at first calving and days open under different censoring scenarios in Iranian Holstein cows.

The present study aimed to investigate the effect of censoring, the situations in which incomplete at the time, out of range, and/or delayed records were considered, in the estimation of genetic parameters for age at first calving (AFC) and days open (DO) in Iranian Holstein cows. The dataset included 281,772 records collected from 1991 to 2019 by the Vahdat Cooperative Company, a pioneer dairy farm in Isfahan Province, the central part of Iran. Five animal models including linear model (LM), penalty model (PM), modified penalty model (MPM), linear-threshold model (LTM), and modified linear-threshold model (MLTM) were used for genetic evaluation of the trait studied. The predictive ability of the models was assessed using cross-validation. The lowest mean square of error and highest r(y,y) were obtained under MLTM for AFC and under LTM for DO, indicating that MLTM and LTM are recommended for genetic evaluation of AFC and DO with censored records in Iranian Holstein cows, respectively. The prediction accuracy of the models for AFC was relatively similar, ranging from 0.46 (under LM) to 0.48 (under PM, LTM, and MLTM). For DO, prediction accuracy values ranged from 0.36 (under LM) to 0.47 (under PM and LTM). The posterior mean for heritability of AFC under MLTM was 0.11. There was no significant difference among posterior means for the heritability of AFC under different models. Therefore, LM is preferred for genetic evaluation of AFC in Iranian Holsteins, and taking censored records into account is unnecessary. The posterior mean for heritability of DO under LTM was 0.09. There were no statistically significant differences among the heritability estimates of DO under LTM, PM, and MLTM. But considering censored records for genetic evaluation of DO affects the estimation of heritability and improved model accuracy for this trait. Therefore, LTM is preferred and recommended for genetic evaluation of DO in Iranian Holsteins.

Genomic predictions for fillet yield and firmness in rainbow trout using reduced-density SNP panels.

BACKGROUND: One of the most important goals for the rainbow trout aquaculture industry is to improve fillet yield and fillet quality. Previously, we showed that a 50 K transcribed-SNP chip can be used to detect quantitative trait loci (QTL) associated with fillet yield and fillet firmness. In this study, data from 1568 fish genotyped for the 50 K transcribed-SNP chip and ~ 774 fish phenotyped for fillet yield and fillet firmness were used in a single-step genomic BLUP (ssGBLUP) model to compute the genomic estimated breeding values (GEBV). In addition, pedigree-based best linear unbiased prediction (PBLUP) was used to calculate traditional, family-based estimated breeding values (EBV). RESULTS: The genomic predictions outperformed the traditional EBV by 35% for fillet yield and 42% for fillet firmness. The predictive ability for fillet yield and fillet firmness was 0.19-0.20 with PBLUP, and 0.27 with ssGBLUP. Additionally, reducing SNP panel densities indicated that using 500-800 SNPs in genomic predictions still provides predictive abilities higher than PBLUP. CONCLUSION: These results suggest that genomic evaluation is a feasible strategy to identify and select fish with superior genetic merit within rainbow trout families, even with low-density SNP panels.

Plasma fibrinogen acts as a predictive factor for pathological complete response to neoadjuvant chemotherapy in breast cancer: a retrospective study of 1004 Chinese breast cancer patients.

BACKGROUND: The aim of this study was to evaluate the relationship between pre-treatment plasma fibrinogen (Fib) level and pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) in breast cancer patients and to assess the role of plasma Fib as a predictive factor. METHODS: Data from 1004 consecutive patients with invasive breast cancer who received NAC and subsequent surgery were retrospectively analysed. Both univariate and multivariate analyses based on logistic regression model were performed to identify clinicopathological factors associated with pCR to NAC. Cox regression model was used to determine the correlation between clinical or pathological parameters and recurrence-free survival (RFS). The Kaplan-Meier method and the log-rank test were applied in the survival analysis. RESULTS: The median value of Fib, rather than other plasma coagulation parameters, was significantly increased in non-pCR patients compared with pCR patients (P = 0.002). Based on the cut-off value estimated by the receiver operating characteristic (ROC) curve analysis, patients were divided into low or high Fib groups (Fib < 3.435 g/L or ≥ 3.435 g/L). Low Fib levels were significantly associated with premenopausal or perimenopausal status (P <  0.001), tumour size ≤5 cm (P = 0.002), and positive hormone receptor status (P = 0.002). After adjusted for other clinicopathological factors in the multivariate logistic regression model, low Fib status was strongly associated with pCR to NAC (OR = 3.038, 95% CI 1.667-5.537, P <  0.001). Survival analysis showed that patients with low Fib levels exhibited better 3-year RFS compared with patients with high Fib levels in the tumour size>5 cm group (77.5% vs 58.4%, log-rank, P = 0.0168). CONCLUSIONS: This study demonstrates that low pre-treatment plasma Fib (Fib < 3.435 g/L) is an independent predictive factor for pCR to NAC in breast cancer patients. Moreover, T3-featured breast cancer patients with lower Fib level exhibit better RFS outcomes after NAC compared with high Fib status.

Role of handgrip strength in predicting new-onset diabetes: findings from the survey of health, ageing and retirement in Europe.

BACKGROUND: Diabetes is a major concern for the global health burden. This study aimed to investigate the relationship between handgrip strength (HGS) and the risk of new-onset diabetes and to compare the predictive abilities between relative HGS and dominant HGS. METHODS: This longitudinal study used data from the Survey of Health, Ageing and Retirement in Europe (SHARE), including 66,100 European participants aged 50 years or older free of diabetes at baseline. The Cox proportional hazard model was used to analyze the relationship between HGS and diabetes, and the Harrell's C index, net reclassification index (NRI), and integrated discrimination improvement (IDI) were calculated to evaluate the predictive abilities of different HGS expressions. RESULTS: There were 5,661 diabetes events occurred during follow-up. Compared with individuals with lowest quartiles, the hazard ratios (95 % confidence intervals) of the 2nd-4th quartiles were 0.88 (0.81-0.94), 0.82 (0.76-0.89) and 0.85 (0.78-0.93) for dominant HGS, and 0.95 (0.88-1.02), 0.82 (0.76-0.89) and 0.60 (0.54-0.67) for relative HGS. After adding dominant HGS to an office-based risk score (including age, gender, body mass index, smoking, and hypertension), the incremental values of the Harrell's C index, NRI, IDI of relative HGS were all slightly higher than those of dominant HGS in both training and validation sets. CONCLUSIONS: Our findings supported that HGS was an independent predictor of new-onset diabetes in the middle-aged and older European population. Moreover, relative HGS exhibited a slightly higher predictive ability than dominant HGS.

Predictive ability of the G8 screening test to determine probable sarcopenia and abnormal comprehensive geriatric assessment in older patients with solid malignancies.

BACKGROUND: Pre-treatment evaluation for sarcopenia is recommended in cancer patients. New screening tests that are less time-consuming and can identify patients who will potentially benefit from geriatric assessment are being developed; the G8 geriatric screening test is one such example. We aimed to investigate whether the G8 screening test can detect probable sarcopenia and is valid and reliable compared to a comprehensive geriatric assessment (CGA) in Turkish older adults with solid cancers. METHODS: We included solid cancer patients referred to a single center. Probable sarcopenia and abnormal CGA were defined as low handgrip strength. Cut-offs for handgrip strength in the Turkish population have been previously determined to be 32 kg for males and 22 kg for females and impairment in at least one of the CGA tests, respectively. The CGA tests comprised KATZ Basic Activities of Daily Living Scale Lawton-Brody Instrumental Activities of Daily Living Scale, Mini-Mental-State Examination Scale, Geriatric Depression Scale-15, and Mini-Nutritional Assessment Short Form. Receiver operating characteristic curve analyses evaluated the test's predictive ability. Intra-rater and inter-rater reliabilities were assessed. RESULTS: The median age of the 76 patients included was 72 (65-91) years. There was a moderate correlation between handgrip strength and the G8 test total score. The sensitivity and specificity of the G8 test to detect probable sarcopenia alone (cut off score = 12.5) were 50 and 92%, respectively (AUC: 0.747; p < 0.001); to determine abnormal CGA plus probable sarcopenia (cut off score = 13) were 93.33 and 86.89%, respectively (AUC: 0.939; p < 0.001); and to detect abnormal CGA alone (cut off score = 14) were 79.63 and 95.45%, respectively (AUC: 0.893; p < 0.001). The G8 test results agreed with those of CGA (κ = 0.638; p < 0.001). Both inter- and intra-rater assessments of G8 scores revealed a strong agreement (Interclass correlation coefficient = 0.979, p < 0.001 and ρ = 0.994, p < 0.001, respectively). CONCLUSIONS: The Turkish version of the G8 test is a good screening tool to detect probable sarcopenia alone and in conjunction with abnormal CGA in older patients with solid malignancies. The G8 screening tool may thus be useful in detecting probable sarcopenia in Turkish older adults with solid cancers.

Evaluation of the efficiency of genomic versus pedigree predictions for growth and wood quality traits in Scots pine.

BACKGROUND: Genomic selection (GS) or genomic prediction is a promising approach for tree breeding to obtain higher genetic gains by shortening time of progeny testing in breeding programs. As proof-of-concept for Scots pine (Pinus sylvestris L.), a genomic prediction study was conducted with 694 individuals representing 183 full-sib families that were genotyped with genotyping-by-sequencing (GBS) and phenotyped for growth and wood quality traits. 8719 SNPs were used to compare different genomic with pedigree prediction models. Additionally, four prediction efficiency methods were used to evaluate the impact of genomic breeding value estimations by assigning diverse ratios of training and validation sets, as well as several subsets of SNP markers. RESULTS: Genomic Best Linear Unbiased Prediction (GBLUP) and Bayesian Ridge Regression (BRR) combined with expectation maximization (EM) imputation algorithm showed slightly higher prediction efficiencies than Pedigree Best Linear Unbiased Prediction (PBLUP) and Bayesian LASSO, with some exceptions. A subset of approximately 6000 SNP markers, was enough to provide similar prediction efficiencies as the full set of 8719 markers. Additionally, prediction efficiencies of genomic models were enough to achieve a higher selection response, that varied between 50-143% higher than the traditional pedigree-based selection. CONCLUSIONS: Although prediction efficiencies were similar for genomic and pedigree models, the relative selection response was doubled for genomic models by assuming that earlier selections can be done at the seedling stage, reducing the progeny testing time, thus shortening the breeding cycle length roughly by 50%.

How Strong the Predictive Ability of the Variable Age and Stroke Severity for Patients Presented Late.

BACKGROUND AND PURPOSE: Age and stroke severity as 2 independent predictors have been included in many well-known prediction models. However, whether the model consisting of the 2 variables derived from early arrival group could bring equal clinical benefit for those patients presented late was unclear. This study aimed to investigate the performance of this transformation. METHODS: We enrolled ischemic stroke patients admitted to our stroke center within 3 days after symptom onset from January 1, 2018 to March 31, 2019.These patients were divided into 2 groups, early arrival group within 6 hours after onset and late arrival group between 6 hours and 3 days. Two multivariate logistic regression models were developed consisting of the variable age and stroke severity. The primary outcome was the unfavorable outcome which defined as modified Rankin Scale score of 3-6. The differences of the performance of the models were compared through 3 aspects (discrimination, calibration, and clinical utility). RESULTS: Five-hundred seventeen ischemic stroke patients were included in our study. There were 258 patients reached in our stroke center within 6 hours while 259 patients were not. The area under the curve were .78 (95% confidence interval .70-.87) for the model developed in the early arrival group and .82 (95% confidence interval .73-.90) for the model developed in the late arrival group respectively. The models calibrated well in the late arrival group. As for clinical utility, the net benefit of the model developed in the early group was only slightly lower than the model developed in the late arrival group. CONCLUSIONS: The prediction model consisting of the variable age and stroke severity derived from the early arrival group patients had the potential to be applied directly in the patients presented late.

Development and Proof-of-Concept Application of Genome-Enabled Selection for Pea Grain Yield under Severe Terminal Drought.

Terminal drought is the main stress limiting pea (Pisum sativum L.) grain yield in Mediterranean environments. This study aimed to investigate genotype × environment (GE) interaction patterns, define a genomic selection (GS) model for yield under severe drought based on single nucleotide polymorphism (SNP) markers from genotyping-by-sequencing, and compare GS with phenotypic selection (PS) and marker-assisted selection (MAS). Some 288 lines belonging to three connected RIL populations were evaluated in a managed-stress (MS) environment of Northern Italy, Marchouch (Morocco), and Alger (Algeria). Intra-environment, cross-environment, and cross-population predictive ability were assessed by Ridge Regression best linear unbiased prediction (rrBLUP) and Bayesian Lasso models. GE interaction was particularly large across moderate-stress and severe-stress environments. In proof-of-concept experiments performed in a MS environment, GS models constructed from MS environment and Marchouch data applied to independent material separated top-performing lines from mid- and bottom-performing ones, and produced actual yield gains similar to PS. The latter result would imply somewhat greater GS efficiency when considering same selection costs, in partial agreement with predicted efficiency results. GS, which exploited drought escape and intrinsic drought tolerance, exhibited 18% greater selection efficiency than MAS (albeit with non-significant difference between selections) and moderate to high cross-population predictive ability. GS can be cost-efficient to raise yields under severe drought.

Genomic selection for non-key traits in radiata pine when the documented pedigree is corrected using DNA marker information.

BACKGROUND: Non-key traits (NKTs) in radiata pine (Pinus radiata D. Don) refer to traits other than growth, wood density and stiffness, but still of interest to breeders. Branch-cluster frequency, stem straightness, external resin bleeding and internal checking are examples of such traits and are targeted for improvement in radiata pine research programmes. Genomic selection can be conducted before the performance of selection candidates is available so that generation intervals can be reduced. Radiata pine is a species with a long generation interval, which if reduced could significantly increase genetic gain per unit of time. The aim of this study was to evaluate the accuracy and predictive ability of genomic selection and its efficiency over traditional forward selection in radiata pine for the following NKTs: branch-cluster frequency, stem straightness, internal checking, and external resin bleeding. RESULTS: Nine hundred and eighty-eight individuals were genotyped using exome capture genotyping by sequencing (GBS) and 67,168 single nucleotide polymorphisms (SNPs) used to develop genomic estimated breeding values (GEBVs) with genomic best linear unbiased prediction (GBLUP). The documented pedigree was corrected using a subset of 704 SNPs. The percentage of trio parentage confirmed was about 49% and about 50% of parents were re-assigned. The accuracy of GEBVs was 0.55-0.75 when using the documented pedigree and 0.61-0.80 when using the SNP-corrected pedigree. A higher percentage of additive genetic variance was explained and a higher predictive ability was observed when using the SNP-corrected pedigree than using the documented pedigree. With the documented pedigree, genomic selection was similar to traditional forward selection when assuming a generation interval of 17 years, but worse than traditional forward selection when assuming a generation interval of 14 years. After the pedigree was corrected, genomic selection led to 37-115% and 13-77% additional genetic gain over traditional forward selection when generation intervals of 17 years and 14 years were assumed, respectively. CONCLUSION: It was concluded that genomic selection with a pedigree corrected by SNP information was an efficient way of improving non-key traits in radiata pine breeding.

Pea genomic selection for Italian environments.

BACKGROUND: A thorough verification of the ability of genomic selection (GS) to predict estimated breeding values for pea (Pisum sativum L.) grain yield is pending. Prediction for different environments (inter-environment prediction) has key importance when breeding for target environments featuring high genotype × environment interaction (GEI). The interest of GS would increase if it could display acceptable prediction accuracies in different environments also for germplasm that was not used in model training (inter-population prediction). RESULTS: Some 306 genotypes belonging to three connected RIL populations derived from paired crosses between elite cultivars were genotyped through genotyping-by-sequencing and phenotyped for grain yield, onset of flowering, lodging susceptibility, seed weight and winter plant survival in three autumn-sown environments of northern or central Italy. The large GEI for grain yield and its pattern (implying larger variation across years than sites mainly due to year-to-year variability for low winter temperatures) encouraged the breeding for wide adaptation. Wider within-population than between-population variation was observed for nearly all traits, supporting GS application to many lines of relatively few elite RIL populations. Bayesian Lasso without structure imputation and 1% maximum genotype missing rate (including 6058 polymorphic SNP markers) was selected for GS modelling after assessing different GS models and data configurations. On average, inter-environment predictive ability using intra-population predictions reached 0.30 for yield, 0.65 for onset of flowering, 0.64 for seed weight, and 0.28 for lodging susceptibility. Using inter-population instead of intra-population predictions reduced the inter-environment predictive ability to 0.19 for grain yield, 0.40 for onset of flowering, 0.28 for seed weight, and 0.22 for lodging susceptibility. A comparison of GS vs phenotypic selection (PS) based on predicted genetic gains per unit time for same selection costs suggested greater efficiency of GS for all traits under various selection scenarios. For yield, the advantage in predicted efficiency of GS over PS was at least 80% using intra-population predictions and 20% using inter-population predictions. A genome-wide association study confirmed the highly polygenic control of most traits. CONCLUSIONS: Genome-enabled predictions can increase the efficiency of pea line selection for wide adaptation to Italian environments relative to phenotypic selection.