Effect of calf stimulation on milk ejection in reindeer (Rangifer tarandus)

<p>The objective of this study was to establish methods for stimulating the milk ejection in reindeer kept for milking purpose. Calves were used to stimulate milk does’ let down. In experiment 1, five does were allowed olfactory, acoustic and visual contact with their calves during milking, whereas four does were milked in isolation. The treatment of the groups was alternated every day during the eight days experiment. Olfactory, acoustic and visual contact with the calf did not influence the doe’s milk yield. The milk yield varied significantly between individual females within treatment (P < 0.01). In experiment 2, the calves were allowed to suckle their mother for a short period (two seconds) prior to milking being initiated. The same alternate design as in experiment 1 with groups consisting of three and two animals respectively was used, and the experiment lasted four days. The pre-suckling stimulation significantly increased the milk ejection measured as milk yield (P < 0.05), and the residual milk after the treatment was negectible. Moreover, the milk ejection varied between individual females within treatment (P < 0.05). We conclude that it is possible to achieve a complete milk removal by machine milking after the does have been pre-stimulated by suckling of calves. Olfactory, acoustic and visual contact with calves during milking failed to influence the milk ejection in this study. However, the results have to be interpreted with caution due to limited sample size.</p><p>Abstract in Norwegian / Sammendrag: Formålet med dette forsøket var å prøve ut ulike metoder for å stimulere nedgivninga av melk hos rein. Kalvene ble tatt i bruk for å stimulere nedgivninga. I forsøk 1 hadde simla lyd-, lukt og synskontakt med kalven mens melkingen pågikk. Vi benyttet et ”switch back design” der fem simler hadde kontakt med kalven under melkingen og fire ble melket uten kontakt. Behandlingen ble byttet om annenhver dag i de åtte dagene forsøket varte. Lyd-, lukt og synskontakt med kalven under melking hadde ingen innvirkning på melkemengden ved maskinmelking. Det var imidlertid individuell variasjon i hvor mye melk man oppnådde hos simlene (P < 0.01). I forsøk 2 lot vi kalvene suge simlene en kort stund før simlene ble melket. Kalven ble sluppet inn til simla og sugingen ble avbrutt etter to sekunder. Deretter ble simla ført inn for maskinmelking. Forsøket varte i fire dager og vi benyttet samme ”switch back design” som i forsøk 1, med grupper bestående av henholdsvis to og tre dyr. Stimuleringa med suging hadde en signifikant innvirkning på nedgivninga. (P < 0.05), og mengden gjenværende melk var minimal. Dessuten var det en signifikant individuell variasjon i melkemengden innen behandlingen (P < 0.05). Vi konkluderer med at det er mulig å oppnå en fullstendig tømming av juret ved maskinmelking dersom simlene først er blitt stimulert med suging av kalven. Lyd-, lukt- og synskontakt med kalven under melking hadde ingen innvirkning på nedgivninga av melk i dette forsøket. Resultatene må imidlertid tolkes med varsomhet siden det statistiske utvalget er begrenset.</p>

Download Full-text

scHinter: imputing dropout events for single-cell RNA-seq data with limited sample size

Bioinformatics ◽

10.1093/bioinformatics/btz627 ◽

2019 ◽

Author(s):

Pengchao Ye ◽

Wenbin Ye ◽

Congting Ye ◽

Shuchao Li ◽

Lishan Ye ◽

...

Keyword(s):

Gene Expression ◽

Sample Size ◽

Single Cell ◽

Wide Spectrum ◽

Dropout Rate ◽

Supplementary Information ◽

Sample Sizes ◽

Limited Sample ◽

Limited Sample Size ◽

The Impact

Abstract Motivation Single-cell RNA-sequencing (scRNA-seq) is fast and becoming a powerful technique for studying dynamic gene regulation at unprecedented resolution. However, scRNA-seq data suffer from problems of extremely high dropout rate and cell-to-cell variability, demanding new methods to recover gene expression loss. Despite the availability of various dropout imputation approaches for scRNA-seq, most studies focus on data with a medium or large number of cells, while few studies have explicitly investigated the differential performance across different sample sizes or the applicability of the approach on small or imbalanced data. It is imperative to develop new imputation approaches with higher generalizability for data with various sample sizes. Results We proposed a method called scHinter for imputing dropout events for scRNA-seq with special emphasis on data with limited sample size. scHinter incorporates a voting-based ensemble distance and leverages the synthetic minority oversampling technique for random interpolation. A hierarchical framework is also embedded in scHinter to increase the reliability of the imputation for small samples. We demonstrated the ability of scHinter to recover gene expression measurements across a wide spectrum of scRNA-seq datasets with varied sample sizes. We comprehensively examined the impact of sample size and cluster number on imputation. Comprehensive evaluation of scHinter across diverse scRNA-seq datasets with imbalanced or limited sample size showed that scHinter achieved higher and more robust performance than competing approaches, including MAGIC, scImpute, SAVER and netSmooth. Availability and implementation Freely available for download at https://github.com/BMILAB/scHinter. Supplementary information Supplementary data are available at Bioinformatics online.

Download Full-text

The predictive performances of random forest models with limited sample size and different species traits

Fisheries Research ◽

10.1016/j.fishres.2020.105534 ◽

2020 ◽

Vol 227 ◽

pp. 105534 ◽

Cited By ~ 3

Author(s):

Jing Luan ◽

Chongliang Zhang ◽

Binduo Xu ◽

Ying Xue ◽

Yiping Ren

Keyword(s):

Random Forest ◽

Sample Size ◽

Species Traits ◽

Limited Sample ◽

Limited Sample Size ◽

Forest Models ◽

Random Forest Models

Download Full-text

Effect of different milking routines on milking-related release of the hormones oxytocin, prolactin and cortisol, and on milk yield and milking performance in Murrah buffaloes

Journal of Dairy Research ◽

10.1017/s0022029904000457 ◽

2005 ◽

Vol 72 (1) ◽

pp. 10-18 ◽

Cited By ~ 11

Author(s):

Chirathalattu S Thomas ◽

Rupert M Bruckmaier ◽

Karin Östensson ◽

Kerstin Svennersten-Sjaunja

Keyword(s):

Milk Yield ◽

Fold Increase ◽

Threshold Level ◽

Milk Composition ◽

Direct Application ◽

Concentration Increase ◽

Milk Ejection ◽

Machine Milking ◽

Murrah Buffaloes ◽

Cortisol Release

Milking-related release of oxytocin, prolactin, and cortisol was studied following three pre-milking treatments. Six Murrah buffaloes were treated with direct application of milking cluster (O), a 1-min pre-stimulation (M), and combined feeding and pre-stimulation (MF). Machine milk yield, stripping yield and milk composition were recorded. Milk ejection occurred significantly earlier with MF than M and O (P<0·05; 2·50, 5·10 and 6·33 min, respectively). In all treatments, milk ejection occurred with small increases >3–5 ng/l in oxytocin concentration. Increase in oxytocin concentration over a threshold level and milk ejection occurred simultaneously and were closely correlated (r=0·83, P<0·05). There was a positive correlation between total time oxytocin concentration remained elevated over threshold levels and machine yield (r=0·86, P<0·05). For treatment O, milk ejection was inhibited during machine milking, while a marked increase in oxytocin occurred during hand stripping (6 and 16 ng/l, respectively). For treatment M, mean oxytocin concentrations remained unchanged during pre-stimulation but increased during subsequent machine milking and hand stripping (6·38, 18·06 and 12·36 ng/l, respectively). For treatment MF, although there was a 3·6-fold increase during pre-stimulation, oxytocin increased by 10-fold and 3-fold during machine milking and hand stripping, and was significant for machine milking (P<0·05, 17·32, 47·86, 18·13 ng/l, respectively). Milk-ejection-related cortisol release was visible only in treatment MF. For treatments O and M, prolactin concentration increased prior to the increase in oxytocin. The stripping yield was higher, and fat content in the stripping yield significantly lower, for treatment O indicating incomplete milking. Thus buffaloes are easily disturbed even by small changes in milking routines.

Download Full-text

A maximum uncertainty LDA-based approach for limited sample size problems — with application to face recognition

Journal of the Brazilian Computer Society ◽

10.1007/bf03192391 ◽

2006 ◽

Vol 12 (2) ◽

pp. 7-18 ◽

Cited By ~ 9

Author(s):

Carlos Eduardo Thomaz ◽

Edson Caoru Kitani ◽

Duncan Fyfe Gillies

Keyword(s):

Face Recognition ◽

Sample Size ◽

Limited Sample ◽

Limited Sample Size

Download Full-text

Prediction in $\mathcal{M}$ -complete Problems with Limited Sample Size

Bayesian Analysis ◽

10.1214/13-ba826 ◽

2013 ◽

Vol 8 (3) ◽

pp. 647-690 ◽

Cited By ~ 7

Author(s):

Jennifer Lynn Clarke ◽

Bertrand Clarke ◽

Chi-Wai Yu

Keyword(s):

Sample Size ◽

Limited Sample ◽

Limited Sample Size ◽

Complete Problems

Download Full-text

Synthetic observations from deep generative models and binary omics data with limited sample size

10.1101/2020.06.11.147058 ◽

2020 ◽

Cited By ~ 1

Author(s):

Jens Nußberger ◽

Frederic Boesel ◽

Stefan Lenz ◽

Harald Binder ◽

Moritz Hess

Keyword(s):

Sample Size ◽

Binary Data ◽

Generative Models ◽

Small Sample ◽

Generative Adversarial Networks ◽

Omics Data ◽

Sample Sizes ◽

Odds Ratios ◽

Limited Sample ◽

Limited Sample Size

AbstractDeep generative models can be trained to represent the joint distribution of data, such as measurements of single nucleotide polymorphisms (SNPs) from several individuals. Subsequently, synthetic observations are obtained by drawing from this distribution. This has been shown to be useful for several tasks, such as removal of noise, imputation, for better understanding underlying patterns, or even exchanging data under privacy constraints. Yet, it is still unclear how well these approaches work with limited sample size. We investigate such settings specifically for binary data, e.g., as relevant when considering SNP measurements, and evaluate three frequently employed generative modeling approaches, variational autoencoders (VAEs), deep Boltzmann machines (DBMs) and generative adversarial networks (GANs). This includes conditional approaches, such as when considering gene expression conditional on SNPs. Recovery of pair-wise odds ratios is considered as a primary performance criterion. For simulated as well as real SNP data, we observe that DBMs generally can recover structure for up to 100 variables with as little as 500 observations, with a tendency of over-estimating odds ratios when not carefully tuned. VAEs generally get the direction and relative strength of pairwise relations right, yet with considerable under-estimation of odds ratios. GANs provide stable results only with larger sample sizes and strong pair-wise relations in the data. Taken together, DBMs and VAEs (in contrast to GANs) appear to be well suited for binary omics data, even at rather small sample sizes. This opens the way for many potential applications where synthetic observations from omics data might be useful.

Download Full-text

Techniques to Produce and Evaluate Realistic Multivariate Synthetic Data

10.1101/2021.10.26.465952 ◽

2021 ◽

Author(s):

John Heine ◽

Erin E.E. Fowler ◽

Anders Berglund ◽

Michael J. Schell ◽

Steven A Eschrich

Keyword(s):

Sample Size ◽

Random Number Generator ◽

Synthetic Data ◽

Principal Component ◽

Limited Sample ◽

Limited Sample Size ◽

Input Sample ◽

Size Condition ◽

Input Variables ◽

Pca Transform

Background: Proper data modeling in biomedical research requires sufficient data for exploration and reproducibility purposes. A limited sample size can inhibit objective performance evaluation. Objective: We are developing a synthetic population (SP) generation technique to address the limited sample size condition. We show how to estimate a multivariate empirical probability density function (pdf) by converting the task to multiple one-dimensional (1D) pdf estimations. Methods: Kernel density estimation (KDE) in 1D was used to construct univariate maps that converted the input variables (X) to normally distributed variables (Y). Principal component analysis (PCA) was used to transform the variables in Y to the uncoupled representation (T), where the univariate pdfs were assumed normal with specified variances. A standard random number generator was used to create synthetic variables with specified variances in T. Applying the inverse PCA transform to the synthetic variables in T produced the SP in Y. Applying the inverse maps produced the respective SP in X. Multiple tests were developed to compare univariate and multivariate pdfs and covariance matrices between the input (sample) and synthetic samples. Three datasets were investigated (n = 667) each with 10 input variables. Results: For all three datasets, both the univariate (in X, Y, and T) and multivariate (in X, Y, and T) tests showed that the univariate and multivariate pdfs from synthetic samples were statistically similar to their pdfs from the respective samples. Application of several tests for multivariate normality indicated that the SPs in Y were approximately normal. Covariance matrix comparisons (in X and Y) also indicated the same similarity. Conclusions: The work demonstrates how to generate multivariate synthetic data that matches the real input data by converting the input into multiple 1D problems. The work also shows that it is possible to convert a multivariate input pdf to a form that approximates a multivariate normal, although the technique is not dependent upon this finding. Further studies are required to evaluate the generalizability of the approach.

Download Full-text