3709 research outputs found
Sort by
Environmental microbiome mapping as a strategy to improve quality and safety in the food industry
Get PDFpeer-reviewedIn food industries, an environmentally-adapted microbiome can colonize the surfaces of equipment and tools and be transferred to the food product or intermediates of production. These complex microbial consortia may include microbial spoilers, pathogens, as well as beneficial microbes.Advances in sequencing technologies and metagenomics provide the opportunity to map the environmental microbiome in food industries at an unprecedented depth, highlighting the importance of the resident microbial communities in influencing food quality and safety, as well as the main factors shaping its composition and activities. However, specific technical issues must be considered. Although microbiome mapping in the food industry has the potential to revolutionize food safety and quality management systems, its application as routine practice is still challenging and technical issues limit the exploitation of the powerful information that can be obtained by the application of such state-of-the-art approaches
Effect of 3 autumn pasture management strategies applied to 2 farm system intensities on the productivity of spring-calving, pasture-based dairy systems
Get PDFpeer-reviewedThe objective of this study was to investigate the effect of altering autumn pasture availability and farm system intensity on the productivity of spring-calving dairy cows during autumn. A total of 144 Holstein-Friesian and Holstein-Friesian × Jersey crossbred dairy cows were randomly assigned to 2 whole farm system (FS) intensities and 3 autumn pasture availability (PA; measured above 3.5 cm) treatments in a 2 × 3 factorial arrangement. The 2 farm systems consisted of a medium intensity (MI: 2.75 cows/ha, target postgrazing sward height of 4.0–4.5 cm) and high intensity system (HI: 3.25 cows/ha, target postgrazing sward height of 3.5–4.0 cm, + 1.8 kg of concentrate dry matter [(DM)/cow per day]. Within each farm system treatment, cows were further subdivided into 3 different PA management strategies: high PA (HPA), medium PA (MPA), and low PA (LPA). The experimental period lasted for 11 wk from September 1 to housing of all animals on November 20 (±2 d) over 3 yr (2017–2019, inclusive). To establish the different average pasture covers for each PA treatment during autumn and in particular at the end of the grazing season, grazing rotation length was extended by +13 and +7 d for HPA and MPA, respectively, beyond that required by LPA (37 d). There were no significant FS × PA interactions for any of the pasture, dry matter intake, or milk production and composition variables analyzed. There were also no differences in pregrazing sward characteristics or sward nutritive value between FS with the exception of daily herbage allowance, which was reduced for HI system (12.2 vs. 14.2 kg of DM/cow). Milk and milk solid yield were greater for HI groups (15.9 and 1.55 kg/cow per day, respectively) compared with MI (15.4 and 1.50 kg/cow per day, respectively). Mean paddock pregrazing herbage mass was significantly higher with increased PA ranging from a mean of 1,297 kg of DM/ha for LPA to 1,718 and 2,111 kg of DM/ha of available pasture for MPA and HPA, respectively. Despite large differences in pregrazing herbage mass, there was no difference in cumulative pasture production and only modest differences in grazing efficiency and sward nutritive value between PA treatments. On average, closing pasture covers were 420, 650, and 870 kg of DM/ha for LPA, MPA, and HPA, respectively, on December 1. In addition to maintaining similar grazing season lengths and achieving big differences in availability of pasture on farm into late autumn, PA treatment had no significant effect on dry matter intake, milk production, and body condition score during the study period. The results of this study indicate that greater cow performance and pasture utilization can be achieved through a greater daily concentrate allocation along with an increased stocking rate. Moreover, the potential to adapt grazing management practices to increase the average autumn pasture cover in intensive grazing systems is highlighted. In addition, a high dependence on high-quality grazed pasture during late autumn can be ensured without compromising grazing season length while also allowing additional pasture to be available for the subsequent spring
Data and code: Beneficial effects of multi-species mixtures on N2O emissions from intensively managed grassland swards
No full textWe provide the data and statistical code that produced the results presented in Cummins et al. (2021). Three files are available to download:
- Excel data file ‘Cummins_etal_2021_ N2O’: contains the annual N2O emissions and N2O emissions intensity (Biomass N and DM yield) measured from sown grassland comprising one to six species within one to three functional groups (grass, legume and herb).
- An accompanying readme file ‘Readme_Cummins_etal_2021_ N2O’: this describes the data and metadata in the Excel data file.
- Statistical code (SAS software) ‘SAScode_for_Cummins_et_al_2021’: SAS code to repeat the analyses as in the final models of the published paper.We provide the data and statistical code that produced the results presented in Cummins et al. (2021). Three files are available to download:
- The data file ‘Cummins_etal_2021_N2O.csv’: contains the annual N2O emissions and N2O emissions intensities (see Readme file for details) measured from sown grassland comprising one to six species within one to three functional groups (grass, legume and herb).
- An accompanying readme file ‘Readme_Cummins_etal_2021_N2O.txt’: this contains the metadata for the data in ‘Cummins_etal_2021_N2O.csv’.
- Statistical code (SAS software) ‘SAScode_for_Cummins_et_al_2021.sas’: SAS code to repeat the analyses as in the final models of the published paper
Goat farm variability affects milk Fourier-transform infrared spectra used for predicting coagulation properties
Get PDFPeer-ReviewedDriven by the large amount of goat milk destined for cheese production, and to pioneer the goat cheese industry, the objective of this study was to assess the effect of farm in predicting goat milk-coagulation and curd-firmness traits via Fourier-transform infrared spectroscopy. Spectra from 452 Sarda goats belonging to 14 farms in central and southeast Sardinia (Italy) were collected. A Bayesian linear regression model was used, estimating all spectral wavelengths' effects simultaneously. Three traditional milk-coagulation properties [rennet coagulation time (min), time to curd firmness of 20 mm (min), and curd firmness 30 min after rennet addition (mm)] and 3 curd-firmness measures modeled over time [rennet coagulation time estimated according to curd firmness change over time (RCTeq), instant curd-firming rate constant, and asymptotical curd firmness] were considered. A stratified cross validation (SCV) was assigned, evaluating each farm separately (validation set; VAL) and keeping the remaining farms to train (calibration set) the statistical model. Moreover, a SCV, where 20% of the goats randomly taken (10 replicates per farm) from the VAL farm entered the calibration set, was also considered (SCV80). To assess model performance, coefficient of determination (R2VAL) and the root mean squared error of validation were recorded. The R2VAL varied between 0.14 and 0.45 (instant curd-firming rate constant and RCTeq, respectively), albeit the standard deviation was approximating half of the mean for all the traits. Although average results of the 2 SCV procedures were similar, in SCV80, the maximum R2VAL increased at about 15% across traits, with the highest observed for time to curd firmness of 20 mm (20%) and the lowest for RCTeq (6%). Further investigation evidenced important variability among farms, with R2VAL for some of them being close to 0. Our work outlined the importance of considering the effect of farm when developing Fourier-transform infrared spectroscopy prediction equations for coagulation and curd-firmness traits in goats.Università degli Studi di Sassar
The effects of cow genetic group on the density of raw whole milk
Get PDFpeer reviewedThe density of milk is dependent upon various factors including temperature, processing conditions, and
animal breed. This study evaluated the effect of different cow genetic groups, Jersey, elite Holstein Friesians
(EHF), and national average Holstein Friesians (NAHF) on the compositional and physicochemical properties
of milk. Approximately 1,040 representative (morning and evening) milk samples (~115 per month during
9 mo) were collected once every 2 wk. Milk composition was determined with a Bentley Dairyspec instrument.
Data were analysed with a mixed linear model that included the fixed effects of sampling month, genetic
group, interaction between month and genetic group and the random effects of cow to account for repeated
measures on the same animal. Milk density was determined using three different analytical approaches –
a portable and a standard desktop density meter and 100 cm3 calibrated glass pycnometers. Milk density was
analysed with the same mixed model as for milk composition but including the analytical method as a fixed effect.
Jersey cows had the greatest mean for fat content (5.69 ± 0.13%), followed by EHF (4.81 ± 0.16%) and NAHF (4.30
± 0.15%). Milk density was significantly higher (1.0313 g/cm³ ± 0.00026, P < 0.05) for the milk of Jersey breed when
compared to the EHF (1.0304 ± 0.00026 g/cm³) and NAHF (1.0303 ± 0.00024 g/cm³) genetic groups. The results from
this study can be used by farmers and dairy processors alike to enhance accuracy when calculating the quantity
and value of milk solids depending upon the genetic merit of the animal/herd, and may also improve milk payment
systems through relating milk solids content and density
Mid infrared spectroscopy and milk quality traits: a data analysis competition at the “International Workshop on Spectroscopy and Chemometrics 2021”
No full textdatasetchemometric data analysis challenge has been arranged during the first edition of the
“International Workshop on Spectroscopy and Chemometrics”, organized by the Vistamilk
SFI Research Centre and held online in April 2021. The aim of the competition was to build a
calibration model in order to predict milk quality traits exploiting the information contained
in mid-infrared spectra only. Three different traits have been provided, presenting heterogeneous
degrees of prediction complexity thus possibly requiring trait-specific modelling
choices. In this paper the different approaches adopted by the participants are outlined and
the insights obtained from the analyses are critically discussed
Factors Affecting the Welfare of Unweaned Dairy Calves Destined for Early Slaughter and Abattoir Animal-Based Indicators Reflecting Their Welfare On-Farm
Get PDFpeer-reviewedIn many dairy industries, but particularly those that are pasture-based and have seasonal calving, “surplus calves,” which are mostly male, are killed at a young age because they are of low value and it is not economically viable to raise them. Such calves are either killed on farm soon after birth or sent for slaughter at an abattoir. In countries where calves are sent for slaughter the age ranges from 3-4 days (New Zealand and Australia; “bobby calves”) to 3-4 weeks (e.g., Ireland); they are not weaned. All calves are at the greatest risk of death in the 1st month of life but when combined with their low value, this makes surplus calves destined for early slaughter (i.e., <1 month of age) particularly vulnerable to poor welfare while on-farm. The welfare of these calves may also be compromised during transport and transit through markets and at the abattoir. There is growing recognition that feedback to farmers of results from animal-based indicators (ABI) of welfare (including health) collected prior to and after slaughter can protect animal welfare. Hence, the risk factors for poor on-farm, in-transit and at-abattoir calf welfare combined with an ante and post mortem (AM/PM) welfare assessment scheme specific to calves <1 month of age are outlined. This scheme would also provide an evidence base with which to identify farms on which such animals are more at risk of poor welfare. The following ABIs, at individual or batch level, are proposed: AM indicators include assessment of age (umbilical maturity), nutritional status (body condition, dehydration), behavioral status (general demeanor, posture, able to and stability while standing and moving, shivering, vocalizations, oral behaviors/cross-sucking, fearfulness, playing), and evidence of disease processes (locomotory ability [lameness], cleanliness/fecal soiling [scour], injuries hairless patches, swellings, wounds], dyspnoea/coughing, nasal/ocular discharge, navel swelling/discharge); PM measures include assessment of feeding adequacy (abomasal contents, milk in rumen, visceral fat reserves) and evidence of disease processes (omphalitis, GIT disorders, peritonitis, abscesses [internal and external], arthritis, septicaemia, and pneumonia). Based on similar models in other species, this information can be used in a positive feedback loop not only to protect and improve calf welfare but also to inform on-farm calf welfare management plans, support industry claims regarding animal welfare and benchmark welfare performance nationally and internationally
Investigating the Effectiveness of Representations Based on Word-Embeddings in Active Learning for Labelling Text Datasets
Get PDFManually labelling large collections of text data is a timeconsuming
and expensive task, but one that is necessary to support machine
learning based on text datasets. Active learning has been shown
to be an effective way to alleviate some of the effort required in utilising
large collections of unlabelled data for machine learning tasks without
needing to fully label them. The representation mechanism used to represent
text documents when performing active learning, however, has a
significant influence on how effective the process will be. While simple
vector representations such as bag-of-words have been shown to be an effective
way to represent documents during active learning, the emergence
of representation mechanisms based on the word embeddings prevalent
in neural network research (e.g. word2vec and transformer based models
like BERT) offer a promising, and as yet not fully explored, alternative.
This paper describes a large-scale evaluation of the effectiveness of different
text representation mechanisms for active learning across 8 datasets
from varied domains. This evaluation shows that using representations
based on modern word embeddings, especially BERT, which have not yet
been widely used in active learning, achieves a significant improvement
over more commonly used vector representations like bag-of-words
BRIAR: Biomass Retrieval in Ireland Using Active Remote Sensing (2014-CCRP-MS.17)
No full textEPA RESEARCH PROGRAMME 2014–2020, Report 305Biomass Retrieval Using Active Remote
Sensing
Hedgerows are a very significant component of the
Irish landscape. They perform multiple functions,
acting as boundary markers, acting as stock-proof
fencing, supporting bio-diversity and controlling run-off.
They function as reservoirs of above-ground biomass
and their potential as carbon sinks was explored in an
earlier study which found that hedgerows potentially
sequester 0.5–2.7tCO2
/ha/year.
The earlier study used light detection and ranging
(lidar) scanning to build 3D models of hedgerows to
successfully estimate biomass, but at the time the
cost–benefit of doing so was poor. However, this
has since changed with the availability of free lidar
sources and the reduced cost of commissioning/
acquiring lidar data. The purpose of the present study
was to examine the use of another active remote
sensing tool, imaging radar, to estimate biomass in
hedgerows.
The study area around Fermoy in County Cork was
field surveyed using new drone technology to collect
data on a sample of hedgerows from which estimates
of biomass could be drawn. These field estimates were
used with new high-resolution TerraSAR-X Staring
Spotlight (TSX-SS) radar imagery to model hedgerows
directly from radar backscatter.
The study found that hedgerow biomass cannot be
derived directly from radar backscatter. There were
a number of reasons for this, such as the hedgerow
biomass density, with an average of 10kg/m2
, being
above the threshold of saturation for radar in the
X-band frequency range. However, other radar
sensors with lower frequencies, and thus higher
saturation limits, do not have the spatial resolution to
map hedgerows.
An alternative method of investigating hedgerow
structure, and thus inferring biomass, interferometry,
is not successful as the level of coherence between
the observations in our dataset was too low to build
a 3D model (i.e. the backscatter from the hedgerow
changed too much between observations).
A new method that examines the cross-sectional
response of the radar return across a hedgerow was
shown to be successful at modelling the relationship
between the width of the backscatter profile and the
width of the hedgerow. However, this too was sensitive
to the orientation of the hedgerow to the sensor.
Therefore, this study shows that radar data does not
seem to be an appropriate technology for estimating
hedgerow properties in Ireland.
In order to estimate the national stock of hedgerow,
the new Prime2 spatial data storage model (OSI,
2014) was applied in conjunction with developed
maps showing the probability of a field boundary being
a stone wall or a hedgerow, to give a new national
estimate for hedgerow length in Ireland of 689,000km.
This estimate is double the frequently quoted figure
of 300,000km because of a much wider definition of
“hedgerow” used in this report.
Net change in hedgerow length was examined using
the aerial photographic records from 1995, 2005 and
2015, along with county-level survey records, showing
that there has been a net removal of hedgerows
between 1995 and 2015 of between 0.16% and 0.3%
per annum, although the rate is much slower in the
latter half of that period.
As X-band radar seems to be inappropriate for
hedgerow evaluation (especially for the obvious case
of the identification of the complete removal of large
hedgerows, for which it is much more expensive
and time-consuming than the detection of hedgerow
removal using aerial photography), the existing
national lidar surveys from the Geological Survey of
Ireland were examined for their appropriateness for
hedgerow evaluation. A digital canopy model derived
from these data successfully estimated heights (mean
and maximum) in the trial test site, with an r
2
value
of 0.79
Gerry Downey: an authentic spectroscopist
Get PDFThis year has seen the retirement of
Gerry Downey from active service with
the Irish National Agriculture and Food
Research Institute, Teagasc1
in Dublin. As
one of Europe’s leading innovative spectroscopic
chemometricians and a great
positive personality to have as a project
partner, we thought it appropriate
to dedicate a column to Gerry’s career,
however embarrassed he may be about
the idea