Layla Cristien de Cássia Miranda Dias’s research while affiliated with Federal University of Viçosa and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (5)


Figure 1 -Examples of the features extracted from the images of pigs.
Figure 2 -Live weight measurements by breed and by group.
Figure 3 -Regression metrics between predicted and observed data throughout the 100 repetitions of train-test partition.
Figure 4 -Linear regression of the observed weight and the weight predicted by random forest regression approach (best performance), with the respective coefficient of determination (R 2 ), root mean square error (RMSE), and mean absolute error (MAE).
Figure 5 -Pearson's correlation among live weight and the features obtained from image processing for back area, back perimeter, back width, and body depth of evaluated pigs.
Comparison of supervised machine learning and variable selection methods for body weight prediction of growth pigs using image processing data
  • Article
  • Full-text available

November 2024

·

54 Reads

Revista Brasileira de Zootecnia

·

Polliany da Costa Santos Oliveira

·

Layla Cristien de Cássia Miranda Dias

·

[...]

·

This research aimed to compare statistical methods (random forest, RIDGE, LASSO, and elastic net regression) for the prediction of body weight in purebred and crossbred pigs reared in Brazil. This prediction was based on dorsal-view images obtained from video image processing. The study involved 69 animals belonging to breeds such as Large White, Piau, Duroc × Large White, and Piau × Large White. The data collection spanned 144 days, with measurements taken at approximately 20-day intervals, totaling eight measurements for each animal throughout their growth stages. Image acquisition was carried out in individual pens using an Intel RealSense Depth D435 digital camera. The features back area, back perimeter, back width, and body depth were extracted from the images. Pearson’s correlation analysis was conducted to assess the relationship between live weight and these features. The dataset was randomly divided into a training dataset (65%) and a test dataset (35%), and model training was performed by five-fold cross-validation balanced according to the growth stage, which was divided into three groups. This procedure was repeated 100 times, and the resulting metrics were taken as the average of the 100 repetitions. Although with a slight difference, the random forest method outperformed the others with the highest average R² value (0.87), as well as the lowest average RMSE (14.32) and average MAE (10.13) values. Consequently, the random forest algorithm proved to be the most effective in predicting body weight. The back area, back width, and back perimeter were the most important variables in the model. 2D image; back area; crossbred pig; penalized regression; precision livestock farming; random forest

Download

Linear regression coefficient estimates (standard error within parentheses) of inbreeding for weight at birth, weight at weaning, and average pre-weaning daily weight gain (ADG)
Estimation of genetic parameters and inbreeding depression in Piau breed pigs

May 2022

·

48 Reads

This study aimed to estimate genetic parameters, effective population size, inbreeding, and inbreeding depression for birth weight, weaning weight, and average pre-weaning daily weight gain (ADG) in Piau breed pigs. We used information from 3841 Piau pigs, and four linear models were fitted in single trait analyzes including or excluding maternal genetic effect, common litter effect, or a combination of these. The models’ adjustments were compared by likelihood ratio test, in which the model that presented the best fit for each trait was used to estimate (co)variance components. The inbreeding depression effect was evaluated using a linear model that included the fixed effects of sex, parity order, contemporary group, and inbreeding coefficient as a fixed covariate. Weight at birth and weaning showed low direct heritabilities (0.08 and 0.05, respectively), while ADG showed moderate heritability (0.20). Weight at birth showed high genetic correlations with weight at weaning (0.90) and with ADG (0.82). Weight at weaning and ADG also showed a high genetic correlation (0.99). There was an inbreeding increase over the generations and a reduction of the effective population size. In the last generation evaluated all animals were inbred, the average inbreeding coefficient was 0.07 and the effective population size was 20.8. It was observed a significant inbreeding effect on ADG, wherein an increase of 1% on the inbreeding coefficient resulted in a decrease of 0.005 grams on ADG. Thus, increase effective population size is mandatory for the inbreeding control and to reduce the loss of variability in this Piau population.


Reconstrução de pedigree com uso de SNPs comuns a diferentes chips comerciais em Gir Leiteiro

Objetivou-se eleger um painel reprodutível e preciso para teste de paternidade em bovinos Gir Leiteiro, usando marcadores SNP comuns a diferentes chips comerciais. Foram utilizados dados genotípicos de 16.205 animais, provenientes de quatro diferentes chips comerciais para bovinos, sendo utilizados para análise apenas os 1.810 marcadores comuns a eles. O pedigree foi reconstruído com o método de Razão de Verossimilhança e uma probabilidade de confiança foi obtida. Houve 8.799 atribuições de paternidade completas (pai e mãe), 15.066 somente de pai e 9.238 somente de mãe, todas com confianças superiores a 99,995%. Com essa metodologia foi possível a identificação de paternidades que não constavam no pedigree de referência. O painel mostrou-se satisfatório para a reconstrução de pedigree, logo, recomenda-se o uso do mesmo para testes de paternidade em bovinos Gir Leiteiro.


Efeito da herdabilidade e arquitetura genética na acurácia da predição genômica de animais cruzados utilizando dados simulados

October 2021

·

5 Reads

Resumo: Neste estudo, objetivou-se avaliar como diferentes herdabilidades e arquiteturas genéticas das características afetam a acurácia de predição genômica dos animais cruzados a partir de informações das linhas puras. Foram simulados quinze diferentes cenários considerando diferentes herdabilidades total (poligênica + QTL): 0,50, 0,30 e 0,10; e diferentes herdabilidade de QTL (h²QTL) correspondente a 0%, 25%, 50%, 75% e 100% da herdabilidade total. As acurácias de predição dos animais cruzados foram calculadas como a correlação de Pearson entre os valores genéticos estimados (GEBVs), preditos pelo método GBLUP, e os valores genéticos verdadeiros (TBVs) dos animais cruzados. A menor acurácia obtida foi 0,2247 quando considerou-se h² = 0,10 e h² QTL = 0% e a maior acurácia obtida foi 0,9418 ao considerar h² = 0,50 e h²QTL = 100%. Nos cenários completamente poligênico (0% de h² de QTL) e com h² de QTL de 25%, a acurácia de predição foi maior para h² de 0,3 em comparação a h² de 0,5. Nos demais cenários a acurácia aumentou com o aumento da herdabilidade. Independente da h², quanto maior a proporção da variância explicada pelo QTL maior a acurácia de predição. Assim, a acurácia da predição de valores genéticos genômicos para animais cruzados utilizando as linhas puras como população de referência é influenciada tanto pela h² quanto pela proporção de variância explicada pelos QTLs.