Skip to main content
Login | Suomeksi | På svenska | In English

Browsing by discipline "Animal Science (Animal Breeding)"

Sort by: Order: Results:

  • Yrjö-Koskinen, Anneli (2013)
    Most of the organic dairy producers use conventional, non-organic animals to renew their breeding stock due to lack of breeding program for organic production. Milk production and food intake is usually lower amongst organic cows. Conventional production favour cows with great milk performance, which is a trait that is not as beneficial in organic production. Organic feed is often less energetic than the conventional, which easily leads to fertility and metabolic problems in organic cows with large genetic milk performance. The adaptation to different environments depends on the interaction between genotype and environment. Because of this, animals that thrive in one environment do not succeed as well in other environments. This research is based on organic dairy breeders’ opinions, experiences and future objectives. In this study I also draw conclusions on whether the Finnish organic dairy production would be in need of its own breeding program. The results of this paper are based on two questionnaires, which were sent to organic dairy producers in the end of year 2011. The first questionnaire was sent to 124 organic dairy producers and the other questionnaire to all who answered the first one. The first part consisted of basic questions about the producers’ opinions, goals and future plans, as well as a ranking of the most desirable traits in the breeding stock. In the second part the producers were asked, among other things, to rank the seven most desirable traits in a pairwise comparison using a scale of 1 – 9. The returned questionnaires were analysed by using the Analytic Hierarchy Process method (AHP) (Saaty and Vargas 2001). The consistencies of the answers were checked by calculating the matrixes consistency ratios using Microsoft Excel. The results of the AHP analysis showed that the organic producers favoured the following qualities, starting with the most favourable trait: udder health, milk production, temperament, fertility, milking abilities, leg health and body structure. Milk production was considered to be the most important trait amongst production qualities. Protein production came in second place and fat production in third and last place. The producers thought udder conformation to be more important than feet and legs and feet and legs to be more important than body size. The conclusion of the analysis is that organic dairy producers favour production traits differently than what is currently weighted in conventional production. According to organic preferences the organic producers should use local breeds. In practice, however, only 6,5% of the Finnish organic producers reported their herds to consist fully of Finn cattle, while 19,5% had a herd consisting of Finn cattle plus one other breed or several other breeds. This research shows that the majority of the Finnish organic dairy producers have similar breeding goals with the conventional production. Since the largest difference between organic and conventional dairy production is to be found in the environmental conditions, it would be beneficial to breed some of the conventional AI-bulls so that they would adapt better to organic conditions.
  • Niemi, Helka (2013)
    Genetic parameters and relationships were estimated for production, herd life, and udder traits for Finnish Ayrshire and Holstein cows. Data consisted of 25,521 Ayrshire cows and 7703 Holstein cows, born from 1994 to 2005. The estimated production traits were first lactation milk yield and lifetime milk yield, estimated herd life traits were total days in milk and number of lactations. The estimated udder traits were fore attachment, fore teat placement, median suspensory, total udder, rear attachment width, rear attachment length, rear teat placement, balance, teat length, teat thickness, and udder depth. Heritabilities and genetic correlations between traits were calculated by REML using the DMU software. Heritabilities for lifetime milk and herd life traits were near zero. Heritability for first lactation milk was 0.25 for Ayrshire and 0.19 for Holstein. The estimates of heritabilities for udder traits were mainly low to moderate ranging from 0.15 to 0.32. Teat length (Ayrshire 0.32; Holstein 0.29) and udder depth (Ayrshire 0.30; Holstein 0.31) had the highest heritabilities. The genetic correlations between first lactation milk and lifetime milk were moderate (Ayrshire 0.37; Holstein 0.47). Teat length and thickness had a negative genetic correlation to lifetime milk and both herd life traits. The highest positive genetic correlations were between udder depth and herd life traits and between total udder and herd life traits in Ayrshire, ranging from 0.35 to 0.37. Direct selection on lifetime and herd life traits would not be effective because of the low heritabilities and, because longer generation intervals are associated with these traits. Alternatively, progress can be made in herd life traits by putting selection emphasis on udder depth and avoiding thick and long teats.
  • Tikkanen, Minna (2014)
    Genetic and phenotypic parameters and relationships for fertility traits, body type traits and production were estimated. The data analyzed included 21,450 Ayrshires. Animals were reared in 2,647 herds, born from 1994 to 2005 and were progeny of 1652 sires. Analyzed fertility traits were days from first service to last insemination and number of inseminations to conception for heifers, and days from calving to first insemination for first parity cows. Production traits were first lactation milk and lifetime milk and body type traits were stature, body depth, chest width, angularity, top line, rump width and rump angle. Variance components and the heritability estimates were calculated by restricted maximum likelihood (REML) method using the DMU software. The heritability estimates of first lactation milk yield and lifetime milk yield were 0.28 and 0.08, re-spectively. The heritabilities for fertility traits in virgin heifers and first lactation cows were low (0.02 – 0.03). Heritability estimates for the type traits varied from 0.10 to 0.43. The largest heritability was found for stature (0.43) and rump width (0.27). The highest positive genetic correlations were angularity with first lactation milk (0.41) and lifetime milk (0.45), and the highest negative genetic correlation was between top line and first lactation milk (-0.33). Chest width and rump angle had a positive genetic correlation to days from calving to first insemination. The significant genetic correlations were heifer fertility with body depth, rump width and rump angle. Estimated correlations between heifer fertility traits and lifetime milk yield were positive and moderate. First parity cow fertility was not related to production traits. Fertility traits and lifetime milk have a low heritability. Progress can be made relatively fast in body traits and first lactation milk by breeding, because of the moderate heritabilities. Selection for some body type traits may cause improvement in production but deterioration in fertility. Declined heifer fertility is associated with high lifetime production. However is not reasonable to put emphasis on poor heifer fertility in selection.
  • Kuru, Johanna (2010)
    Kirjolohi (Oncorhynchus mykiss) on tärkein Suomessa viljeltävä ruokakalalaji, jonka taloudellisen merkityksen vuoksi Riista- ja kalatalouden tutkimuslaitos ja Maa- ja elintarviketalouden tutkimuskeskus ylläpitävät kirjolohen valintajalostusohjelmaa (JALO-valintaohjelma). Kirjolohen jalostusarvostelussa on mukana kasvua, sukukypsyysikää, lihanväriä ja ulkomuotoa kuvaavia ominaisuuksia. Lisäksi valintaohjelmassa tarkkaillaan muotovirheiden yleisyyttä sekä kaihin esiintymistä jalostuspopulaatiossa. Tämän tutkimuksen tarkoituksena oli selvittää löytyykö kirjolohelta kvantitatiivisen ominaisuuden lokuksia (QTL:ia) fileen väri - ja kalan koko-ominaisuuksille. Tutkimuksessa käytetty F2-polven aineisto koostui neljästä risteytysperheestä, joissa P-polven isät olivat amerikkalaista kantaa ja emät JALO-kantaa. QTL-kartoitusta käyttämällä pyrittiin selvittämään löytyykö kirjolohen ruumiinkoko- ja fileen väriominaisuuksista suurivaikutteisia lokuksia, joihin kytkeytyneiden DNA-merkkien avulla kalamateriaalia pystyttäisiin tulevaisuudessa parantamaan merkkiavusteisella valinnalla. Geneettisinä merkkeinä käytettiin useilta eri lohilajeilta löydettyjä mikrosatelliitteja. F2-polven yksilöiden tutkimuksessa käytetyt ominaisuudet mitattiin kolmannen kasvukauden jälkeen Tervossa syksyllä 2006. Tutkimukseen valittiin kuusi ominaisuutta, jotka voitiin jaotella kalan kokoon ja fileen väriin liittyviin ominaisuuksiin. Koko-ominaisuuksia olivat kalan pituus (mm), paino (g) ja filepaino (kahden fileen paino grammoina) ja fileen väriin liittyviä ominaisuuksia roche, valoisuutta kuvaava lave ja värikylläisyys eli cave. Roche-arvo kuvaa visuaalisesti roche-värikortilla määritettyä fileen väriä. Fileenvärin valoisuus ja värikylläisyys mitattiin kolmesti fileen kyljestä spektrofotometrillä ja lopullinen arvo muodostui kolmen mittauksen keskiarvosta. Mittaustuloksista poistettiin ennen QTL-analyysia sukupuolen, sukukypsyyden ja kaihin vaikutus. Tilastolliseen testaukseen käytettiin varianssianalyysia, jolla analysoitiin onko tutkittavissa ominaisuuksissa tietyn DNA-merkin eri genotyyppien välillä tilastollisesti merkitseviä eroja. Jokaisessa tutkittavassa ominaisuudessa löydettiin useita suurivaikutteisia lokuksia, jotka alittivat tutkimuskohtaisesti valitun riskitason. Kalan kokoon vaikuttavia QTL-alueita, jotka vaikuttivat sekä pituuteen, painoon että filepainoon löydettiin yhdeksästä eri kytkentäryhmästä. Yhtään pelkästään pituuteen liittyvää suurivaikutteista lokusta ei havaittu, kuten ei myöskään pituuteen ja painoon tai pituuteen ja filepainoon liittyvää QTL:sta. Painoon ja filepainoon liittyviä QTL-alueita löydettiin kymmenestä kytkentäryhmästä. Filepaino oli kalan koko-ominaisuuksista ainoa, jolle löytyi pelkästään kyseiseen ominaisuuteen vaikuttavat QTL:t kahdesta kytkentäryhmästä. Fileen väriin vaikuttavia QTL-alueita löytyi useita. Viidestä kytkentäryhmästä löydettiin QTL, joka vaikutti sekä rocheen, fileen valoisuuteen että värikylläisyyteen. Rocheen ja valoisuuteen vaikuttavia QTL-alueita löydettiin kahdesta kytkentäryhmästä. Valoisuuteen ja värikylläisyyteen vaikuttava suurivaikutteinen lokus löydettiin yhdestä kytkentäryhmästä. Jokaiselle fileen värin muuttujista löydettiin vain kyseiseen ominaisuuteen vaikuttava QTL. Rochelle ja valoisuudelle löydettiin kummallekin yksi QTL ja värikylläisyydelle kaksi vain kyseiseen ominaisuuteen vaikuttavaa QTL:sta. Tämän tutkimuksen tulosten perusteella merkkiavusteisella valinnalla on mahdollista parantaa tutkimuksessa mukana olleita kalan kokoon – ja fileen väriin liittyviä ominaisuuksia, sillä näihin ominaisuuksiin kytkeytyneitä DNA-merkkejä voitaisiin käyttää yhtenä valintakriteerinä parhaiden jalostusyksilöiden valinnassa.
  • Mikkola, Lea (2013)
    In the last two decades, a variety of studies have introduced the idea that there are significant differences between levels of variance components and heritabilities, between breeds of beef cattle. In this study the following carcass traits were observed for these differences: slaughter weight, carcass class and carcass fatness. The breeds used for this evaluation were hereford, charolais, aberdeen angus, limousin, simmental and highland cattle. Cattle breeds may differ by their physiological traits and breeds are often divided into groups by their original continents. One of the objectives of this study was to analyse the level of such differences and their impact on the genetic improvement of carcass traits. To successfully define these matters it is essential to determine the genetic parameters of carcass traits for each breed separately. After assessing the levels of genetic parameters for each breed in question, the significance of the differences between genetic variances was evaluated. Two models were compared: an unconstrained model, in which the breeds were seen as different traits, with their own variances and a constrained model in which the breeds were also as separate traits but with one joint variance. The results suggest that the unconstrained model is to be preferred over the constrained model. The differences between the levels of variances and heritabilities were apparent between breeds. For highland cattle the results were only indicative due to the limited amount of data acquired. These differences between the breeds’ genetic parameters may affect the prediction of breeding values severely. Therefore, it may be noted in the light of this study that the best breeding values might be obtained by assessing the breeds separately.
  • Häkli, Katja (2013)
    The objective of this study was to estimate genetic parameters of minks Aleutian Disease (AD) infections and investigate the genetic relationships between Aleutian Disease infection and fertility traits. The research data was obtained from the Finnish Fur Breeders Association. The data had information from one mink farm in years 2006-2010. The data contained 27 753 one year old female minks. Only those females who had more than three kits were tested. There were 18 478 tested minks of which 9.2 % had AD. The pedigree data had 41 573 animals. In this study AD infection, pregnancy and felicity were binary traits. AD inf. 1 = sick, 2 = healthy. If female had kit(s) or aborted or lost her kit(s) after birth, pregnancy (PREG) = 2, otherwise = 1. If female lost her kits after birth, felicity (FEL) = 1. If litter size were at least one, FEL = 2. The heritability estimates and variance components (REML) were calculated using multi trait animal model and DMU-software. Heritability estimates for the studied traits were low: 0.07 for AD infection, 0.07 for PREG, 0.04 for FEL and 0.08 for litter size. The heritability estimate for AD was smaller than the common litter variance for the trait. Genetic correlations between AD inf. and FEL, and AD inf. and litter size were positive. The genetic correlations between AD inf. and FEL were 0.447 ± 0.132 and between AD inf. and litter size 0.290 ± 0.108. The results suggest that resistance for AD inf. has genetic variation. Although the heritability estimate for the trait was low, it can be affected by selection. Genetic correlations between AD inf. and fertility traits were favourable. Selection for larger litter size could increase mink resistance for AD infection.
  • Lehto, Emilia (2012)
    The aim of this study was to investigate the genetic relationships between conformation and performance traits in Finnish Warmblood riding horse population. The research material included the RHQT (riding horse quality test) results from the years 1996-2010 and the pedigree data for the horses. The RHQT results were received from the Finnish Breeders Association and the pedigree data was received from the Finnish trotting and breeding association’s horse register. The RHQT results consisted of 640 horses and the pedigree data consisted of 4 458 horses. The REML estimates of (co)variance components were calculated by VCE6 using multi trait model. Altogether there were six trait groups, which consisted of 27 different traits. Because of the large number of traits, the analyses were done in smaller subsets with two trait groups at the same time. The heritability estimates for the studied traits were quite high and the standard errors were small. The genetic correlations within each trait group were strong and positive. Thus it could be concluded that same genes or groups of genes affect all the traits within the trait group. The other explanation for this could be that in reality the judges have evaluated the same traits as different traits. The genetic correlations between conformation traits and all gait traits and dressage traits indicate that conformation can be used indirectly in the breeding of gait and dressage traits. Furthermore, it could be concluded that the genetic relationships between conformation and show-jumping traits are quite low. The genetic correlations between gait traits and show-jumping traits were low. The genetic correlations between gait traits and dressage traits were positive and strong. These strong and positive correlations indicate that the judged traits are actually same or very similar. From the results of this study can be concluded that all three different tests for show-jumping traits are actually measuring the same traits. Because of this one test would be adequate to produce the information of horses quality, capacity and breeding value. Moreover it can be concluded that there are genetic relationships between conformation and performance in Finnish Warmblood riding horse population. These discovered relationships can be used when breeding traits that have low heritabilities, such as the mechanic of movement in the walk or rideability. Furthermore, these results can be used to improve the judging methods in RHQT and other performance tests.
  • Valkonen, Merja (2012)
    The main objective of this study was to estimate genetic parameters to aggression- and fear-related traits in Rottweiler-breed. We also determined which kind of variables could be useful in genetic studies. We used a questionnaire which was sent to Rottweiler owners. Genetic parameters were estimated from a data including responses from owners of 510 dogs. Analysis were carried out with Restricted Maximum Likelihood method for three different variables: owner impression, behaviour components and single behaviour guestions. Heritability estimates for different types of aggression and fear were low or moderate. Most common aggression was dog-directed aggression while owner-directer aggression was most rare. Aggressive behaviour was divided into three genetic complex: social unsureness, aggression in encounter situations and sharpness. The widest complex, social unsureness, had highest heritability estimate. Dog’s aggression in social situations was related to fear and unsureness. Owner impression –guestions and behaviour components could be useful when studying personality traits because they measure a wider complex than just a single behaviour. In this study social dogs had less aggressive and fearful behaviour. It could be possible to reduce aggression- and fear-related behaviour problems by raising dog’s sociality by breeding. This study showed that owner impressions could be used to measure dog’s sociality.
  • Wallén, Sini (2012)
    It has been challenging to genetically improve reproduction traits with the resources of the conventional breeding because of the low heritability of these traits. In Finland genetic improvement of the reproduction traits hasn’t either meet the expectations entirely. Gene markers are one possibility to make genetic improvement of the reproduction traits more effective. With an association analysis it is possible to find those gene markers which have statistically significant effect on the trait. The purpose of this study was to identify SNP markers associated with reproduction traits in the Finnish Yorkshire pig breed. Other purpose of this study was to find out if both Finnish pig breeds have certain chromosomal regions associated with same reproduction traits. In this study one goal was also to discover genes which are placed near statistically significant SNP markers. Under this study were nine female reproduction traits which are total number of piglets born in first and later parities, number of stillborn piglets in first and later parities, piglet mortality between birth and weaning in first and later parities, age at first farrowing, first farrowing interval and second farrowing interval. Prior to SNP association analysis, unstandardized EBVs were deregressed and corresponding weights were calculated. The association between SNP markers and deregressed EBV was studied using a mixed linear model, for each SNP separately. The model included a fixed SNP effect and a random polygenic effect. The analyses were performed using the AI-REML method. Statistical significance of the associations was based on Bonferroni-corrected P-values. In this study 20 statistically significant (P-value ? 2,0E-06) associations on the reproduction traits were observed. One of these SNP markers (P-value = 8,39E-08, on chromosome 7) is associated in total number of piglets born in first parity, two markers (P-values = 2,77E-07 and 1,91E-06, on chromosomes 7 and 1) in total number of piglets born in later parities, one marker (P-value = 9,72E-08, on chromosome 8) in number of stillborn piglets in first parity, three markers (P-value = 2,12E-07 and next two P-value was 7,01E-07, on chromosome 13) in piglet mortality between birth and weaning in first parity, 11 markers (P-values between 1,74E-06 and 3,89E-08, on chromosome 7) in age at first farrowing and two markers (P-values = 4,03E-07, on chromosome 7) in second farrowing interval. In this study statistically suggestive (P-value ? 4,0E-06) associations were also observed on chromosomes 1 and 8. These are associated in total number of piglets born in later parities and number of stillborn piglets in first parity. If these chromosomal regions found in this study are confirmed, these SNP markers will be valuable in the national breeding program through their use in marker-assisted selection. With MAS it’s possible to select pigs which have favourable SNP markers as for reproduction traits. In the long run genetic improvement of the reproduction traits is going to accelerate. There is still need of more extensive population analysis in order to discover genes associated reproduction traits and to estimate those gene effects.
  • Riipinen, Mirja (2014)
    Suomessa poikii vuosittain n. 70 000 liharotuista lehmää ja hiehoa. Vaikeaksi luokiteltavia poikimisia on noin 1500 ja vasikoita menetetään satoja. Vaikea poikiminen tuottaa eläimelle kipua ja rasitusta sekä aiheuttaa kustannuksia karjanomistajalle. Yleisesti lihakarjan eläinvalinnan tavoitteena on tuottaa kohtuullisella syntymäpainolla hyvin kasvavia teuraseläimiä. Tässä tutkimuksessa oli suomalaisen charolais-lihakarjan havaintoaineistoa vuosien 2002–2012 väliltä. Mukana oli 27 271 syntymäpainohavaintoa, 7717 vuodenpainohavaintoa ja 2774 poikimavaikeushavaintoa, joista estimoitiin perinnölliset tunnusluvut. AI–REML-menetelmää käyttäen laskelmat toteutettiin monen ominaisuuden maternaalimallilla DMU-ohjelmalla. Kiinteinä tekijöinä olivat sukupuoli, vasikan sukupuoli, syntymäkausi, poikimaikäluokka ja karja–vuosi-tekijä. Tavoitteena oli saada geneettistet korrelaatiot kolmelle tutkimuksessa mukana olevalle ominaisuudelle, jolloin eläinvalintaa voitaisiin tehostaa poikimavaikeuksien vähentämiseksi. Charolais-rodusta otettiin mukaan ensimmäinen poikiminen, koska oletuksena oli, että silloin havainnoissa olisi vaihtelua. Tuloksista havaittiin, että syntymäpainon ja poikimavaikeuden välillä oli melko voimakas positiivinen geneeettinen korrelaatio. Vuodenpainon ja poikimavaikeuden välillä oli lievä positiivinen geneettinen korrelaatio. Tavoitteena ollutta kolmen ominaisuuden maternaalimallia tuotettuja perinnöllisiä tunnuslukuja ei saatu poikimavaikeuden osalta. Poikimavaikeuden maternaalinen periytymisaste oli lähellä nollaa, joten menetelmä ei tuottanut siitä tulosta. Tulokset vastasivat hyvin aikaisempien tutkimusten tuloksia, joten sillä perusteella niitä voidaan pitää luotettavina. Samalta ajanjaksolta estimoitiin geneettiset trendit kaikille kolmelle ominaisuudelle. Tulos osoitti, että syntymäpainon suunta oli lievästi nouseva, vuodenpaino oli kasvanut ja poikimavaikeudet vähentyneet. Geneettisten trendien perusteella eläinvalinnat olivat tuottaneet haluttua tulosta.
  • Laurén, Hannele (2013)
    Ilmiötä, jossa kahden tai useamman kytkeytyneen lokuksen alleelitaajuuksissa tapahtuu muutoksia valinnan kohdistuessa yhden lokuksen yhteen alleeliin, kutsutaan liftausvaikutukseksi. Liftausvaikutuksen merkitystä arvioitiin kahdessa naudan genomin kohdassa, kromosomissa 1 polled-lokuksen läheisyydessä ja kromosomissa 20 kasvuhormonireseptorigeenin läheisyydessä. Sarvellisuuteen vaikuttavan polled-lokuksen lähellä tapahtunutta liftausvaikutusta arvioitiin suomenkarjassa ja maidontuotanto-ominaisuuksiin vaikuttavan QTL:n lähellä tapahtunutta liftausvaikutusta intensiivisen jalostuksen kohteena olevilla roduilla, ayrshirellä ja holsteinilla. Liftausvaikutuksen arviointiin tarvitaan muita populaatioita tai rotuja, joissa vastaavaa liftausvaikutusta ei oleteta esiintyvän. Tätä tarkoitusta varten tutkimuksessa analysoitiin samat kromosomialueet myös viidellä muulla rodulla. Lisäksi tutkittavien kromosomien muuntelun ja muun genomin muuntelun vertailuun käytettiin muissa kuin tutkittavissa kromosomeissa 1 ja 20 genotyypitettyjä mikrosatelliittimerkkejä. Tutkimuksessa oli mukana kolme erillistä aineistoa (BTA1, BTA20 ja vertailuaineisto) yhteensä 41 eri mikrosatelliittilokusta. Tutkimusaineistot käsittivät 10 eri nautarotua (ayrshire, itäsuomenkarja, istobeninkarja, holstein, jakutiankarja, jersey, khomogor, länsisuomenkarja, pohjoissuomenkarja ja ukrainanharmaakarja), yhteensä 377 eläintä. Kaikista aineistoista laskettiin roduittain ja lokuksittain populaatiogeneettiset tunnusluvut (alleelien lukumäärä, alleelirikkaus, alleelitaajuus, ƒ-arvo, heterotsygotia-asteet). Hardy-Weinbergin tasapainotilaa sekä lokusten parittaista kytkentäepätasapainoa tarkasteltiin roduittain jokaisessa kolmessa aineistossa. Liftausvaikutuksen havaittavuutta analysoitiin kahdella eri menetelmällä. F-arvon regressioon perustuva testaus suoritettiin Wiener ym. (2003) artikkelin mukaisesti. Lokusten neutraalisuutta analysoitiin Beaumontin ja Nicholsin (1996) esittämän menetelmän avulla, käyttäen FDIST2-ohjelmaa. Vertailuaineistosta laskettujen populaatiogeneettisten tunnuslukujen perusteella voidaan sanoa, että populaatiot ja niistä tehdyt otokset ovat liftausvaikutuksen tutkimukseen sopivia. Populaatioissa ei esiinny sukusiitosta satunnaisen pariutumismallin olettamusta enempää. Aineistojen alleelirikkauden ja geenidiversiteetin keskiarvot olivat keskenään hyvin samansuuntaisia. Lokusten parittaisessa kytkentäepätasapainossa aineistot BTA1 ja BTA20 poikkesivat selvästi vertailuaineistosta. Aineistoissa BTA1 ja BTA20 kytkentäepätasapainossa olevia lokuspareja oli merkittävästi enemmän kuin vertailuaineistossa. Kummankaan aineiston, BTA1:n tai BTA20:n, muuntelun määrän muutokset eivät tulleet selkeästi näkyviin F-arvon regressioon tai Beaumontin ja Nicholsin (1996) menetelmään perustuvissa testeissä. Tutkimuksessa löytyi viitteitä liftausvaikutuksesta, mutta havainnot eivät ylittäneet tilastollisen merkitsevyyden rajaa näissä aineistoissa. Ilmiön testaukseen olisi kenties tuonut lisää todistusvoimaa aineiston selkeämpi jakautuminen voimakkaammin erisuuntaan jalostettuihin nautarotuihin, esim. maidontuotanto vs. lihantuotanto.