Prediction of non-muscle invasive bladder cancer outcomes assessed by innovative multimarker prognostic models.

BACKGROUND: We adapted Bayesian statistical learning strategies to the prognosis field to investigate if genome-wide common SNP improve the prediction ability of clinico-pathological prognosticators and applied it to non-muscle invasive bladder cancer (NMIBC) patients. METHODS: Adapted Bayesian sequential threshold models in combination with LASSO were applied to consider the time-to-event and the censoring nature of data. We studied 822 NMIBC patients followed-up >10 years. The study outcomes were time-to-first-recurrence and time-to-progression. The predictive ability of the models including up to 171,304 SNP and/or 6 clinico-pathological prognosticators was evaluated using AUC-ROC and determination coefficient. RESULTS: Clinico-pathological prognosticators explained a larger proportion of the time-to-first-recurrence (3.1 %) and time-to-progression (5.4 %) phenotypic variances than SNPs (1 and 0.01 %, respectively). Adding SNPs to the clinico-pathological-parameters model slightly improved the prediction of time-to-first-recurrence (up to 4 %). The prediction of time-to-progression using both clinico-pathological prognosticators and SNP did not improve. Heritability (h (2)) of both outcomes was <1 % in NMIBC. CONCLUSIONS: We adapted a Bayesian statistical learning method to deal with a large number of parameters in prognostic studies. Common SNPs showed a limited role in predicting NMIBC outcomes yielding a very low heritability for both outcomes. We report for the first time a heritability estimate for a disease outcome. Our method can be extended to other disease models.

Product versus additive model for studying artificial insemination results in several livestock populations.

Two models can be used for studying binary results of AI. The additive threshold model proposes an underlying variable as summing the environmental and genetic effects from the 2 individuals involved in the mating, and the product threshold model assumes that the conditional probability of AI success is the product of the probabilities of success of 2 unobserved binary phenotypes (one is the male fertility; the other is the female fertility). The purpose of this paper is to compare the predictive ability of the product and the additive threshold models for studying AI results and to compare results obtained with the 2 models in 3 different species: cattle, sheep, and rabbits. Results showed that the predictive ability of the product model is similar to the additive model in sheep and rabbits but worst in cattle (percentage of wrong prediction = 42, 27, and 35% in the additive model; 43, 28, and 47% in the product model in sheep, rabbits, and cattle, respectively). Even when the 2 models have similar performance, they differed in their EBV (for instance, Pearson correlation between EBV predicted with the 2 models = 0.46 in sheep for male fertility). The product model can determine which sex is responsible for an AI failure. In sheep, the female was the responsible in 94% of the cases and male in 2% of them; in rabbits, the female was the responsible in 54% of the cases and the male in 39% of them. Different estimates of probabilities for male and female fertility success obtained with the product model in the 3 species suggest that male and female fertilities behave differently depending on the species and the uniqueness of the data sets. Although product model seems to provide additional information in the fertility process, further research is needed to understand the worst performance of the product model in cattle.

An application of change-point recursive models to the relationship between litter size and number of stillborns in pigs.

We developed and implemented change-point recursive models and compared them with a linear recursive model and a standard mixed model (SMM), in the scope of the relationship between litter size (LS) and number of stillborns (NSB) in pigs. The proposed approach allows us to estimate the point of change in multiple-segment modeling of a nonlinear relationship between phenotypes. We applied the procedure to a data set provided by a commercial Large White selection nucleus. The data file consisted of LS and NSB records of 4,462 parities. The results of the analysis clearly identified the location of the change points between different structural regression coefficients. The magnitude of these coefficients increased with LS, indicating an increasing incidence of LS on the NSB ratio. However, posterior distributions of correlations were similar across subpopulations (defined by the change points on LS), except for those between residuals. The heritability estimates of NSB did not present differences between recursive models. Nevertheless, these heritabilities were greater than those obtained for SMM (0.05) with a posterior probability of 85%. These results suggest a nonlinear relationship between LS and NSB, which supports the adequacy of a change-point recursive model for its analysis. Furthermore, the results from model comparisons support the use of recursive models. However, the adequacy of the different recursive models depended on the criteria used: the linear recursive model was preferred on account of its smallest deviance value, whereas nonlinear recursive models provided a better fit and predictive ability based on the cross-validation approach.

A comparison between different survival and threshold models with an application to piglet preweaning survival in a dry-cured ham-producing crossbred line.

Different approaches for predicting genetic merit of piglet preweaning survival were compared using proportional hazard, threshold (TM), and sequential threshold (STM) models. Data were from 13,924 crossbred piglets (1,347 litters), born from 2000 to 2006, and originated by mating 189 Large White C21 Gorzagri boars to 328 Large White-derived crossbred sows. A frailty proportional hazard model was fitted assuming 2 different baseline hazard functions (Cox and Weibull time-dependent model) and including sire and nursed litter as random effects. The TM and STM included the same effects as considered in the proportional hazard model. Model fitting was evaluated in terms of goodness of fit and predictive ability. The goodness-of-fit was evaluated using the local weighted regression and the mean squared error, whereas the predictive ability was assessed by using a cross-validation procedure. Estimated sire variances for piglet preweaning mortality were low, and heritability ranged from 0.04 to 0.06. All 4 models led to similar ranking of sires. Results suggest that STM may be preferred to the other models for genetic evaluation of piglet preweaning survival, both for its better predictive ability and its easier interpretation. Further, STM is computationally less demanding than survival models and allows for estimating different variance components from birth up to weaning.

Inbreeding depression on female fertility and calving ease in Spanish dairy cattle.

Inbreeding depression on female fertility and calving ease in Spanish dairy cattle was studied by the traditional inbreeding coefficient (F) and an alternative measurement indicating the inbreeding rate (DeltaF) for each animal. Data included records from 49,497 and 62,134 cows for fertility and calving ease, respectively. Both inbreeding measurements were included separately in the routine genetic evaluation models for number of insemination to conception (sequential threshold animal model) and calving ease (sire-maternal grandsire threshold model). The F was included in the model as a categorical effect, whereas DeltaF was included as a linear covariate. Inbred cows showed impaired fertility and tended to have more difficult calvings than low or noninbred cows. Pregnancy rate decreased by 1.68% on average for cows with F from 6.25 to 12.5%. This amount of inbreeding, however, did not seem to increase dystocia incidence. Inbreeding depression was larger for F greater than 12.5%. Cows with F greater than 25% had lower pregnancy rate and higher dystocia rate (-6.37 and 1.67%, respectively) than low or noninbred cows. The DeltaF had a significant effect on female fertility. A DeltaF = 0.01, corresponding to an inbreeding coefficient of 5.62% for the average equivalent generations in the data used (5.68), lowered pregnancy rate by 1.5%. However, the posterior estimate for the effect of DeltaF on calving ease was not significantly different from zero. Although similar patterns were found with both F and DeltaF, the latter detected a lowered pregnancy rate at an equivalent F, probably because it may consider the known depth of the pedigree. The inbreeding rate might be an alternative choice to measure inbreeding depression.

Impact of calving ease on functional longevity and herd amortization costs in Basque Holsteins using survival analysis.

The aim of this study was to analyze the impact of calving ease (CE) on functional longevity of Basque Holsteins, using a Weibull proportional hazards model. The data considered for the analysis were 53,353 calving records from 25,810 Holstein cows distributed across 781 herds and sired by 746 bulls. The effects included in the statistical model were age at first calving, stage of lactation, interaction between year and season of calving, 305-d adjusted milk yield, CE, herd, and sire. Calving ease was considered as a time-dependent covariate and, as was the case for the rest of covariates included in the model, had a significant effect on functional longevity. Calvings needing assistance or surgery increased culling risk by 18%, when compared with unassisted calvings. The effect of CE on length of productive life in primiparous and multiparous cows was also investigated. A second analysis was performed replacing the CE effect with the interaction between parity and CE to evaluate the effect of CE in primiparous and multiparous cows. An increase in calving difficulty had a greater impact on culling during first lactations than in subsequent ones. Therefore, difficult calvings, mainly at first parities, had a high impact on herd amortization costs, increasing them by 10% in relation to easy calvings. Therefore, calving difficulty should be avoided as much as possible, especially in primiparous cows, to avoid reduction of profitability.

Consequences of selection for yield traits on calving ease performance.

The impact of different breeding goals on the genetic response for calving ease (CE) and yield traits was studied in the Basque Holstein cattle population. The economic value for CE was estimated with a bioeconomic model, using Basque production and market circumstances and taking into account the categorical nature of CE. The economic value for CE was -18.03 euros/cow per calving interval per liability unit. This value was relatively insensitive to changes in the market price of animals but was more sensitive to changes in the incidence of dystocia. Records from parities between 1995 and 2002 were used for the estimation of genetic parameters for yield (actual milk, fat, and protein yield) and CE using a multivariate model. Linear sire models for yield traits and a threshold sire-maternal grandsire model for CE were used. A Holstein population was simulated to determine the consequences of including CE in the breeding goal. Three selection strategies were considered: 1) selection only on yield traits, 2) selection on yield and direct CE (DCE), and 3) selection on yield, DCE, and maternal CE (MCE). Selection on yield traits only resulted in a slight reduction of dystocia. Selection strategies in which DCE or DCE and MCE were included in the breeding goal did not improve the genetic response for DCE and MCE obtained with the first selection strategy. Genetic responses were also calculated using the 2.5th, 50th, and 97.5th percentiles of posterior densities of genetic correlations between DCE and MCE and yield traits. Because responses in CE were sensitive to deviations in estimates of genetic parameters, the inclusion of CE in the monitoring scheme is recommended. Genetic evaluation of bulls for CE is of considerable value because it provides farmers with the opportunity to use assortative matings of sires with favorable estimated breeding values for DCE to primiparous cows.

Analysis of fertility and dystocia in Holsteins using recursive models to handle censored and categorical data.

A method based on the analysis of recursive multiple-trait models was used to 1) estimate genetic and phenotypic relationships of calving ease (CE) with fertility traits and 2) analyze whether dystocia negatively affects reproductive performance in the next reproductive cycle. Data were collected from 1995 through 2002, and contained 33,532 records of CE and reproductive data of 17,558 Holstein cows distributed across 560 herds in official milk recording from the Basque Country Autonomous Community (Spain). The following fertility traits were considered: days open (DO), days to first service, number of services per pregnancy (NINS), and outcome of first insemination (OFI). Four bivariate sire and sire-maternal grandsire models were used for the analyses. Censoring existed in DO (26.49% of the data) and NINS (12.22% of the data) because of cows having been sold or culled before reaching the next parturition. To avoid bias, a data augmentation technique was applied to censored data. Threshold models were used for CE and OFI. To consider that CE affects fertility and the genetic determination of CE and fertility traits, recursive models were applied, which simultaneously considered CE as a fixed effect on fertility performance and the existence of a genetic correlation between CE and fertility traits. The effects of CE score 3 (difficult birth) with respect to score 1 (no problem) for days to first service, DO, NINS, and OFI were 8 d, 31 d, 0.5 services, and -12% success at first insemination, respectively. These results showed poorer fertility after dystocia. Genetic correlations between genetic effects of fertility traits and CE were close to zero, except for the genetic correlations between direct effects of DO and CE, which were positive, moderate, and statistically different from 0 (0.47 +/- 0.24), showing that genes associated with difficult births also reduce reproductive success.