Small Ruminant Research 138 (2016) 111

Contents lists available at ScienceDirect

Small Ruminant Research

journal homepage: www.elsevier.com/locate/smallrumres

Fatty acid synthase (FASN) gene polymorphism and early lactation

milk fat composition in Xinong Saanen goats
Abiel Berhane Haile a , Wei Zhang a , Wei Wang a , Dikun Yang a , Yongqing Yi a , Jun Luo a,b,
a
Shaanxi Key Laboratory of Molecular Biology for Agriculture, College of Animal Science and Technology, Northwest A&F University, Yangling, Shaanxi,
712100, PR China
b
Xinong Saanen Dairy Goat Farm, College of Animal Science and Technology, Northwest A&F University, Yangling, Shaanxi, 712100, PR China

a r t i c l e

i n f o

Article history:
Received 20 December 2014
Received in revised form 5 March 2016
Accepted 22 March 2016
Available online 26 March 2016
Keywords:
Xinong Saanen
FASN
SNP marker
Haplotypes
Milk fatty acid

a b s t r a c t
Fatty Acid Synthase (FASN) is a multifunctional protein, catalyzing the de novo fatty acid (FA) synthesis,
its high mRNA expression in goat mammary epithelial cells (GMECs) substantiated by the medium to
high heritability of milk fat (MF), makes it a candidate gene for association analysis with milk fat prole.
The study aimed to develop markers of Capra hircus FASN to improve healthfulness of goat milk FAs and
to investigate early lactation MF prole. 300 milk samples were collected, 30 days postpartum, from 300
Saanen does, from 2 herds sired by 73 bucks, and were analyzed with gas-chromatography. Linear mixed
models, that considered the effects of herd, herd test day (HTD), parity, doe, sire and allele and haplotype
effects were used to investigate the association of 3 FASN SNPs (911 C/T intron1 SNP1, 852 A/G intron2
SNP2 and 14420 T/C exon37 SNP3) or two haplotypes (H1, C-A-T and H2, T-G-C, constructed from the
three SNPs) in separate models, with 31 individual FAs, 8 FA groups, and 7 FA indices, by SPSS 20, REML
function. A single test per SNP and haplotype was performed to avoid the effect of multiple testing. H1,
SNP1 and SNP2 were most desirable for milk healthfulness because they were signicantly associated
with lower concentrations of myristic acid (C14:0) and palmitic acid (C16:0) and higher concentrations
of linoleic acid (C18:2 n-6, cis). Herd followed by HTD and parity were the predominant factors affecting
FA levels, indicating an effect of nutrition and management. Lower levels of de novo synthesized FAs
but higher levels of C18:1 cis-9 were observed, indicating mobilization of body fat reserves. Although
considered as de novo, butyric acid levels were consistent throughout lactation and showed negative
correlations with other de novo FAs, suggesting other sources of origin. Majority of the odd chain fatty
acids (OCFAs) originate from rumen bacteria, hence they may reect rumen microbial activity. In our
study this was conrmed by the observed positive correlations of all OCFAs with rumenic CLA (through
its precursor C18:1 trans-11) and other trans-FAs, which are end products of biohydrogenation. The SNP
markers developed will assist in marker assisted selection, and the early lactation milk FA analysis will
help in deciphering factors of variation.
2016 Elsevier B.V. All rights reserved.

1. Introduction
Provision of sufcient and healthy food to the ever-growing population of the world is the primary goal of agriculture. However,
despite the enormous gains in food production, food quality and
its health connotations has been a big public health concern. Food
borne illnesses due to saturated fat and cholesterol contained in
meat and dairy products of livestock has been claiming many lives

Corresponding author at: Shaanxi Key Laboratory of Molecular Biology for

Agriculture, College of Animal Science and Technology, Northwest A&F University,
Yangling, Shaanxi, 712100, PR China.
E-mail addresses: luojun6566@sina.com, luojun@nwsuaf.edu.cn (J. Luo).
http://dx.doi.org/10.1016/j.smallrumres.2016.03.025
0921-4488/ 2016 Elsevier B.V. All rights reserved.

globally. Ruminant milk contributes a signicant amount of this

saturated fat consumed by humans, particularly C16:0 (palmitic
acid) along with C14:0 (myristic acid) and C12:0 (lauric acid) have
been associated with cardiovascular diseases (Von Eckardstein,
2006). Hence, there has been a growing interest in identifying the
causal factors for fatty acid variation, to implement appropriate
measures to modulate the fraction of the components.
Milk fat yield and composition can be considerably modied
by genetic, physiological and nutritional means (Chilliard et al.,
2003). Nutritional modication of ruminant milk & meat fatty
acids holds the predominant plausible means, although it is limited
by rumen biohydrogenation and introduction of bacterial origin
fatty acids (Vlaeminck et al., 2004; Kim et al., 2007; Bou et al.,
2009; Decker and Park, 2010; Jung et al., 2010; Nakov, 2010).

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

Biohydrogenation of unsaturated fatty acids in the rumen is about

85% for polyunsaturated fatty acids (Murphy et al., 1987). Hence,
there is higher concentration of saturated fatty acids in tissues
and milk of ruminants than the levels in the feed. High presence
of saturated fatty acids in human diet is a high health risk factor.
Nonetheless, medium to high heritability of milk fat composition
give a niche for improving milk fat quantity and quality genetically
(Bauman et al., 2004; Bastin et al., 2011).
The quality and quantity of milk fat is inuenced by many
candidate genes and transcription factors. A number of molecular
pathways and intricate gene networks are involved in its regulation
and secretion (Bionaz and Loor, 2008). Each individual gene has a
limited role in modulating the expression, in which conventional
animal breeding is less effective as the small effects of each gene are
hard to detect. (Hayes and Goddard, 2001; Dekkers, 2004; Rhode,
2013). However dynamics of science has shifted the paradigm of
conventional animal breeding to Marker assisted selection or/and
Genomic selection, in which organisms can be selected based on
their DNA (Hayes and Goddard, 2001; Nakaya and Isobe, 2012).
Theory and simulation studies agree this method of selection has a
potential to expedite genetic improvement (Raadsma, 2004). Candidate gene SNP association analysis is one method of identifying
variants associated with phenotypic traits, aiding phenotype-based
selection. A non synonymous mutation in the candidate genes can
affect protein structure and function. Besides, synonymous mutations and intronic variants have been observed affecting traits,
explained by their linkage disequilibrium with other loci, which are
causal variants for the observed association (Luo and Wu, 2001; Du
et al., 2007). Moreover, signicant amounts of miRNAs (microRNAs)
are mapped within introns of host genes. MiRNAs are short non
coding RNAs which can down-regulate gene expression by silencing specic target mRNAs (Barik, 2008; Horie et al., 2009; Berillo
et al., 2013). Hence, variations in the intron could affect transcription of some miRNAs which may have an impact on their regulatory
activity.
Milk from a dairy goat provides a healthy source of nutrients for
human consumption (Sanz Ceballos et al., 2009; Silanikove et al.,
2010; Abbas et al., 2014). Small fat globules and high content of
short and medium chain fatty acids ease digestion and absorption into the enterocytes, respectively, adding value for human
consumption speeding up metabolism (Hui and Chandan, 2007).
Besides, short and medium chain fatty acids have no effect on
cholesterol concentrations (McNamara and Hillers, 1986). Goat
milk is rich in vaccenic and conjugated linoleic acids, which have
therapeutic effects on cancer, artherogenesis, diabetes, obesity,
immune modulation and bone formation (Parodi, 1999; Belury,
2002; Bauman et al., 2004; Tyagi and Kathirvelan, 2006; Whigham
et al., 2007). Moreover goat milk is naturally homogenized; articial homogenization has a negative impact of damaging the fat
globule cells, which leads to the release of superoxide Xanthine
Oxidase (free radical) which is one of the agents for DNA mutation
(Enig, 2003).
Due to its central role in de novo fatty acid synthesis, we hypothesized some of the observed variation in milk fat prole might be
caused by polymorphism in FASN gene. FASN is a multifunctional
enzyme that catalyzes the de novo fatty acid synthesis in cells. In
ruminants it is involved in synthesizing all short-chain, medium
chain and half of the amount of palmitic acids (C16:0) (Zhu et al.,
2014). In Xinong dairy goats the gene is located in chromosome
19 and encodes the homodimeric multifunctional enzyme. It has
8217 bp comprised of 42 exons, an ORF of 7545 bp, and 5 and 3 UTR regions of 88 bp and 584 bp respectively with the start codon
located in the 2nd exon and the stop codon is in 42nd exon. It
encompasses seven active functional domains, which participate
in all the processes of fatty acid synthesis. It encodes 3514 amino

acids with an approximate molecular weight (MW) of 273.8 kDa

(Zhu et al., 2014).
The major precursor of de novo fatty acid synthesis in ruminants
is acetate and to a lesser extent beta hydroxybutyrate and propionate. Acetate is activated by the acetyl CoA short chain gene family
to convert it to acetyl CoA. Acetyle CoA carboxylase (ACACA) catalyzes a two-step reaction, carboxylation of acetyle coA to malonyl
coA. Malonyl CoA is a substrate for the action of FASN to catalyze
the sequence of reactions to synthesis palmitate, adding two carbon atoms acetyl-CoA to the growing chain of fatty acid (Hillgartner
et al., 1995). The thioesterase catalytic domain of FAS (thioesterase
I) catalyzes the chain-termination reaction and results in the formation of C16:0 fatty acid (palmitic acid), the principal product
of the FAS reaction in ruminants and monogastric animals (Barber
et al., 1997). Fatty acids with 16 or more carbons cannot be further
elongated by the mammary gland and are susceptible to hydrolysis by thioesterase I, thus short- and medium-chain fatty acids are
synthesized within the mammary epithelial cells de novo. So, the
proportion of these fatty acids in milk reect the de novo fatty acid
contribution to total milk fat. The majority of short and medium
chain (C4:0C14:0) fatty acids and approximately one-half of the
C16:0 fatty acid in milk are derived from de novo fatty acid synthesis (Palmquist, 2006).
Higher levels of FASN gene mRNA expression were scored in
adipose, intestine and mammary gland tissues followed by lungs,
stomach, muscle, spleen, liver and kidney. Negligible levels of
expression were recorded in heart (Zhu et al., 2014). This indicates
FASN genes role as one of the major genes which affects lipogenesis
in goat mammary epithelial cells.
Several association analysis of FASN gene in bovine identied
causal SNPs with signicant associations with milk fatty acid proles (Morris et al., 2007; Ordovas et al., 2008; Schennink et al.,
2009; Matsuhashi et al., 2010; Matsumoto et al., 2012; Yun et al.,
2012). GWAS and QTL mapping studies identied FASN gene in
bovine chromosome 19, as a candidate gene responsible for some
of the variations in milk fat percentage and yield (Nakov, 2010;
Bouwman et al., 2011; Li et al., 2014). The association of short
and medium chain saturated fatty acids with BTA 19 around FASN
region in the study, reinforce FASNs key role in the de novo biosynthesis of short and medium chain fatty acids.
In sheep a SNP in FASN 5 UTR was found to be signicantly associated with milk fat yield. The predicted mRNA secondary structure
suggested one of the alleles caused a stable folding (Sanz et al.,
2013). To our knowledge any association studies between FASN
gene and goat milk prole have not yet been done. Considering the
genes central role in milk fat synthesis it is necessary to scan the
region for variants associated with the indicated traits.

2. Materials and methods

2.1. Animals and sampling
488 Saanen does belonging to two herds owned by Northwest
Agricultural & Forestry university farm and Qianyang county farm
were genotyped for association analysis. The Northwest A&F university animals does belong to 22 half sib families generated by AI
(Articial insemination) ranging from 1 to 25 does per sire and the
Qianyang farm goats belongs to 51 half sib families generated by
controlled buck breeding, ranging from 2 to 12 does per sire. Samples of 5 mL of blood were collected from the does using acid citrate
dextrose (ACD) as an anticoagulant around 30 days postpartum. It
was immediately stored in 80 C refrigerator until use. The data
recorded were animal, milk yield (at milk sampling), herd, days in
milk and parity.

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

The laboratory analysis and analytical determination of 39 FA

were carried out in Shaanxi key Laboratory of Molecular Biology
for Agriculture, College of Animal Science and Technology, Northwest A&F University. The feeding regimen of northwest A&F farm
was corn silage about 3 kg, alfalfa hay about 0.6 kg and concentrate about 1.2 kg (in a ratio of 5:1:2 respectively). The mean total
quantity of feed was 4.8 kg a day. Soybean meal and rapeseed meal
comprised 19 and 6% of the concentrate ration. To our knowledge
the feeding regimen of second farm was similar.
2.2. Milk sampling, extraction of fatty acids and GC analysis
Milk samples were collected from January 2, 2014 to February
30, 2014 from 300 animals, around 30 days postpartum (early lactation). Milk samples of 45 mL of 2/3 morning and 1/3 evening were
pooled for each test day for fatty acid analysis by GC (gas chromatography). All the samples were analyzed by one machine to
minimize effect of analyzer.
The milk fat was extracted by double centrifugation method
as described by Luna et al. (2005). The fatty acids were extracted
with hexane:isopropanol according to Hara and Radin (1978), and
fatty acid methyl esters (FAME) were prepared by base-catalyzed
(KOCH3 ) trans methylation according to Chouinard et al. (1999).
Fatty acid methyl esters were quantied by gas chromatography
using an Agilent 7890A gas chromatograph (Agilent Technologies, Palo Alto, CA) equipped with a fused-silica capillary column
(SPTM -2560; 100m 0.25 mm 0.2 m lm thickness; SupelcoTM
Inc., Bellefonte, PA), and a ame-ionization detector with hydrogen as a carrier gas. Initial oven temperature was 70 C, it was
increased by 13 C/min to 175 C and held for 27 min, again it was
increased by 4 C/min to 215 C and held for 27 min. Inlet and detector temperatures were 250 C with a 100:1 split ratio. Gas constant
ows were held at hydrogen carrier 1 mL/min and detector hydrogen 25 mL/min, airow 400 mL/min, and nitrogen plus carrier at
40 mL/min.
FAME were identied from the retention time in the gas chromatographic analysis, using pure methyl ester standards (Supelco
37Comp. FAME Mix 10 mg/ml in CH2 CL2 , USA 2012) and linoleic
acid, conjugated methyl esters (Sigma, USA 2014) for 37 fatty acids
and 2CLA, respectively. The peak areas in the chromatogram were
lined by one person to minimize effect of analyzer, and were calculated and normalized using response factors. The individual FA
contents were expressed as weight percentages (g/100 g of total
FA). In this work, 45 FA measurements including 37 individual FA,
3 groups of FA based on saturation level (SFA, MUFA and PUFA);
and 5 FA indexes, C14, C16 and C18 desaturation index and 6/3
were carried out.
2.3. DNA extraction, PCR condition, SNP detection and DNA
sequencing
Genomic DNA samples were extracted from the blood by the
standard phenol/chloroform method and following BloodGen Mini
kit. Polymerase chain reaction (PCR) primers covering 42 exons and
intron fragments were designed using Primer v5.0 software (PREMIER Biosoft International, California, USA) based on the nucleotide
sequence of Caprine (NCBI Reference Sequence: NC 022311.1)
and bovine (NCBI Reference Sequence: AC 000176.1) FASN gene
obtained from NCBI. All the primers used are summarized in the
supplemen. 30 animals were used for DNA pooling for initial polymorphism screening of FASN gene. The PCR amplications were
performed using Bio rad S1000TM Thermal Cycler (Berkeley California, USA) in a 20 L reaction volume containing 50 ng of genomic
DNA, 10 L of 10 multi pcr mix (Applied biosystems, USA) and
0.8 pmol of each primer. The PCR products were sequenced by
GenScript (Nanjing) Co., Ltd., and Invitrogen, Shanghai, China. The

sequences were analyzed for identication of SNPs using DNAstar

Lasergene SeqMan pro, software. Untranslated regions, introns and
exons with a positive result for SNPs were genotyped by PCR-RFLP.

2.4. DNA genotyping, allele frequencies and haplotypes

construction
Allele and genotype frequencies, Chi square test (conducted
by comparing the expected and observed genotype frequencies)
to examine departure from the Hardy Weinberg equilibrium, and
linkage disequilibrium (LD) between SNP pairs (D and r2 ) were
computed and haplotypes were constructed collectively using
haploview software.
(http://www.broadinstitute.org/scientic-community/
science/programs/medical-andpopulation-genetics/haploview/
haploview).

2.5. Statistical analysis

From 488 genotyped animals, 300 animals milk fat was
obtained for association analysis. Linear mixed model was used to
perform statistical analysis with SPSS 20 statistical software. Allele
and haplotype substitution models were used to test the association with individual milk fatty acids, fatty acid groups and fatty acid
indexes. The models were as follows.
Model 1. To observe the allele substitution effect of the individual SNPs
Y ijklmn = + Hi + TDji + Ski + Pl + Dmi + SNPnm n + eijklmn
Yijklmn refers to each of 46 dependent variables comprising 31
individual fatty acids, 8 fatty acid groups and 7 fatty acid indices.
refers to the general mean. The factor Hi refers the xed effect
of herd (there were two levels). The independent variable TDji represents to the xed effect of test day within the herd (there were
three levels for herd 1 and one level for herd 2). Ski refers to the
random effect of buck siring doe within the ith herd (There were
22 and 51 sires in Xinong and Qianyang herds, respectively). The
factor Pl refers to the xed effect of parity of doe; there were 3 categories for parity of does (Category 1 (parity 13); category 2 (parity
46); category 3 (parity 79)). Parity and age factors are closely correlated in dairy goat, so generally only one of them is included in
the mathematical models. The factor Dmi refers the random genetic
effect of the mth animal (m = 1300). The factor SNPnm represents
frequency of major allele coded as 0, 1 or 2 carried by the mth animal; n represents the partial regression coefcient, which is allele
substitution effect of the mth SNP, and the factor e represents the
random residual for each observation. A single test per SNP was
performed to avoid the effects of multiple testing.
Model 2. Haplotype substitution effect
Y ijklmn = + Hi + TDji + Ski + Pl + Dmi + n Hnm + eijklmn
For haplotype substitution analysis the same model, but the SNP
effects SNPnm were substituted by haplotype effects n Hnm . Hnm
representing the haplotype effect tted as a covariate coded as 0, 1
or 2 for the number of each haplotypes present in the mth doe. n
is the partial regression coefcient that is haplotypes substitution
effect for the nth haplotype as a deviation from the effect of most
frequent haplotypes set to zero. Similarly a single test per haplotype
was performed to contain type-I error.

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

Table 1
Summary of milk fatty acid composition.
Fatty acid components

Mean

Std. Deviation

Fatty acid components

Mean

Std. Deviation

FAT wt%
C4:0 wt%
C6:0 wt%
C8:0 wt%
C10:0 wt%
C11:0 wt%
C12:0 wt%
C13:0 wt%
C14:0 wt%
C14:1 wt%
C15:0 wt%
C16:0 wt%
C16:1 wt%
C17:0 wt%
C17:1 wt%
C18:0 wt%
C18:1 trans-9 wt%
C18:1 cis-9 wt%
C18:2 n-6 trans wt%
C18:2 n-6 cis wt%
C20:0 wt%
C18:3 n-6 wt%
C18:3 n-3 wt%
C20:1 n-9 wt%
C18:2cis-9, trans-11 CLA wt%a
C18:2 trans-10, cis-12 CLA wt%

3.51
1.91
1.95
2.28
7.27
0.22
3.17
0.12
8.27
0.12
0.77
24.41
0.78
0.93
0.48
12.22
0.54
29.76
0.21
2.76
0.31
0.06
0.28
0.08
0.34
0.04

1.32
0.31
0.35
0.60
2.21
0.17
1.06
0.05
1.47
0.07
0.17
2.30
0.23
0.37
0.18
2.21
0.18
4.94
0.05
0.39
0.14
0.05
0.11
0.10
0.11
0.04

C22:0 wt%
C20:3 n-6 wt%
C20:3 n-3 wt%
C20:5 n-3 wt%
C22:6 n-3 wt%
MCFA wt%b
SFA wt%
MUFA wt%
PUFA wt%
UFA wt%
SFA/UFA w
AI wt%c
OMEGA3 wt%
OMEGA6 wt%
OMEGA6OMEGA3
C14 DES-INDEXd
C16 DES-INDEXe
C18 DES-INDEXf
ELONGATION-INDEXg
GSF wt%h

0.054
0.035
0.28
0.043
0.038
13.427
64.002
31.827
4.151
35.978
1.841
1.754
0.550
3.180
7.210
0.039
0.031
0.712
0.625
11.500

0.028
0.034
0.150
0.030
0.033
2.979
5.294
5.116
0.517
5.270
0.433
0.496
0.211
0.421
4.994
0.023
0.009
0.041
0.049
3.088

a
b
c
d
e
f
g
h

CLA (Conjugated linoleic acid).

MCFA (C6:0C15:0).
AI (Atherogenic index = [12:0 + 4(14:0) + 16:0]/(SMUFA + SPUFA).
C14 Desaturation index = 14:1/(14:0 + 14:1).
C16 Desaturation index = 16:1/(16:0 + 16:1).
C18 Desaturation index = 18:1/(18:0 + 18:1).
Elongation index = (18:0 + 18:1)/(16:0 + 16:1 + 18:0 + 18:1).
Goat specic fatty acid (C6:0, C8:0 and C10:0).

3. Results

C18 desaturation indices and elongation indices (0.80, and 0.62,

respectively) were high.

3.1. Descriptive statistics

3.2. Observed correlations of individual and groups of fatty acids
The descriptive statistics of individual fatty acids, fatty acid
groups and fatty acid indices are presented in Table 1. Mean
value of milk fat was 3.51%. We have identied 39 specic
FAs but weve considered 30 comprising 99.66% of the fatty
acids, the remaining were eliminated from the analysis for concentrations were very low. The milk atherogenic index was
1.75 (Table 1) (Ulbricht and Southgate, 1991). SFA was the
major component of FA composition (64 wt%) followed by UFA
(35.97 wt%) primarily composed of MUFA (31.82 wt%). The PUFA
concentration in milk was only 4.15 wt%. Goat specic fatty
acids (GSF) represented only 11.5%. The range observed among
milk samples were: atherogenic index (0.693.92), concentrations of SFA (45.8878.3 wt%), UFA (21.6954.11 wt%), MUFA
(18.6649.58 wt%), PUFA (1.576.32 wt%), and SFA/UFA ratio
(0.853.61). The major FA in milk was oleic acid (18:1cis-9;
29.76 wt%), followed by palmitic acid (16:0; 24.41 wt%), stearic
acid (18:0; 12.21 wt%), myristic (14:0; 8.27 wt%), and capric
acid (C10:0; 7.26 wt%) acids. The lowest FA levels were generally observed on very long chain fatty acids, with the least
scored by tricosanoic acid (C23:0, 0.004 wt%). Most of them were
excluded from the analysis due to very low concentrations. All-cis5,8,11,14,17-eicosapentaenoic acid (C20:5 n3) averaged 0.043 wt%
and all-cis-11,14,17-eicosatrienoic acid (C23:3 n3) 0.28 wt% and
were detected only in one farm. Very long chain fatty acids were
mainly present in one farm. The lowest (0.056 wt%) of fatty acid
measures, from long chain fatty acids below 20 carbon chain was
observed in gamma linolenic acid (C18:3 n-6). C14 and C16 desaturation indices were low, 0.04 and 0.03, respectively, whereas the

The correlation results of the fatty acids are presented in Tables

25. In the current study individual fatty acids were highly correlated to their respective groups of fatty acids, the higher the fraction
they represent the higher the correlation (Data not shown). SCFA
(C4:0) showed negative correlation (P = 0.01) patterns with all de
novo fatty acids (C6:0C15:0) ranging from 0.17 (C6:0) to 0.37
(C12:0) contrary to the presumed similar source of origin. SCFA also
showed positive correlation (P = 0.01) with C16:0 (0.27), C18:1 cis-9
(0.18), C20:0 (0.14) and C18:3 n6 (0.12), while negative correlation
(P = 0.01) was observed with rumenic CLA (C18:2, cis-9, trans-11)
(0.16). There were signicant correlations within medium even
chain fatty acids (C6C14) (P = 0.01) ranging from 0.54 (C6:0 Vs
C14:0) to 0.95 (C6:0 and C8:0). Degree of correlations were related
to chain length differences, the closest the chain length the highest the degree of correlation and vice versa. Palmitic acid (C16:0)
showed a positive correlation (P = 0.01) with myristic and myristeloic acid (C14:0 (0.49) and C14:1 (0.32) which dissipates with
increasing chain length differences, and showed weak negative
correlation (p = 0.05) with caproic acid C6:0 (0.13) and caprylic
acid C8:0 (0.12) (when the chain length differences increases the
positive correlation decreases to negative correlations). Medium
even chain fatty acids (C6:0C14:0) showed negative correlations
(P = 0.01) with long chain fatty acids, including stearic acid (C18:0),
oleic acid (C18:1 cis-9) and linoleic acid (C18:2 n-6 cis) ranging from
0.15 (C12:0 versus C18:2 n-6, trans) to 0.9 (C10 versus C18:1 cis9), and positive correlation were observed with C18:1 trans 9 fatty
acid (C6:0 (0.25), C8:0 (0.25) and C10:0 (0.23)). Fatty acids C10
(P = 0.05) and C12 (P = 0.01) showed positive correlations of 0.16

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

Table 2
Pearson bivariate correlation of de novo fatty acids.

C4:0
C6:0
C8:0
C10:0
C11:0
C12:0
C13:0
C14:0
C14:1
C15:0
C16:0
C16:1
a
b

FAT

C4:0

C6:0

C8:0

C10:0

C11:0

C12:0

C13:0

C14:0

C141:0

C15:0

C16:0

0.050
0.155a
0.186a
0.191a
0.141
0.213b
0.086
0.113
0.009
0.178a
0.046
0.123

0.174b
0.343b
0.359b
0.170b
0.371b
0.313b
0.180b
0.012
0.219b
0.265b
0.019

0.946b
0.837b
0.296b
0.648b
0.447b
0.539b
0.029
0.152b
0.131a
0.457b

0.934b
0.345b
0.794b
0.587b
0.617b
0.028
0.223b
0.128a
0.407b

0.375b
0.909b
0.721b
0.776b
0.142a
0.374b
0.071
0.396b

0.381b
0.362b
0.315b
0.119a
0.157b
0.022
0.094

0.828b
0.818b
0.257b
0.380b
0.087
0.289b

0.635b
0.339b
0.495b
0.081
0.130a

0.360b
0.413b
0.489b
0.307b

0.151b
0.324b
0.291b

0.252b
0.219b

0.036

Correlation is signicant at the 0.05 level (2-tailed).

Correlation is signicant at the 0.01 level (2-tailed).

and 0.17 with CLA. All odd chain fatty acids (C11:0, C13:0, C15:0 and
C17:0) showed positive correlations (P = 0.01) with medium chain
fatty acids (C6:0 to C14:0) ranging from 0.15 (C15:0 versus C6:0)
to 0.83 (C13:0 versus C12:0). Interestingly they had shown positive
correlations (P = 0.01) with rumenic CLA ranging from 0.21 (C15:0)
to 0.44 (C17:0), with C18:1 trans-9 ranging from 0.15 (C15:0) to
0.25 (C17:0) and with C18:2 n-6 trans ranging from 0.13 (C11:0) to
0.21 (C17:0), while negative correlations were observed with C18:1
cis-9 and C18:2 n-6 cis. CLA showed negative correlation (P = 0.01)
of 0.37 with stearic acid (C18:0) and positive correlations (P = 0.01)
with C18:1 trans-9 (0.37), C18:2 n-6 trans (0.38) and C18:2 n-6 cis
(0.28). Moreover, as shown in Table 6, herd had affected majority
of the fatty acid signicantly followed by parity and HTD.

3.3. Genotyping results

DNA Sequences of pooled DNA revealed 11 genetic variants,
including 911 C/T Intron 1 (Fig. 1), 852 A/G Intron 2 (Fig. 2), 14420
T/C Exon 37 (Fig. 3), which were genotyped and used in our association analysis. The remaining SNPs revealed by sequencing but not
genotyped are presented in the supplementary of the manuscript
(2695 A/G Intron 4, 3158 C/T Exon 6, 3763 A/G Intron 8, 5195 T/A
Intron 10, 7754 G/C Intron 17, 8076 C/G Exon 19, 13733 G/C Intron
34, 16100 G/A Intron 41).
All the does were genotyped for SNPs 911 C/T Intron 1(SNP
1), 852 A/G Intron 2(SNP 2), 14420 T/C Exon 37 (SNP 3) by PCRRFLP using the restriction endonucleases, HindIII, AvaI and Tth111I,
respectively. The digestion products were separated on a 2.0%
agarose gel. In SNP 911 C/T Intron 1, the T allele (uncut) was indicated by a band at 687 bp and the C allele (cut) resulted in a band at
225 bp TT (687 bp), CC (225 bp, 462 bp), CT (687 bp, 225 bp, 462 bp)
(Fig. 1). In SNP 852 A/G Intron 2, the A allele (uncut) was indicated by
a band at 558 bp and the G allele (cut) at 319 bp band AA (558 bp);
GG (319 bp, 239 bp); AG (558 bp, 319 bp, 239 bp) (Fig. 2). In SNP
14420 T/C Exon 37, the C allele (uncut) was indicated by a band at
787 bp and the T allele (cut) was indicated by a band at 439 bp CC
(787 bp); TT (439 bp, 348 bp); TC (787 bp, 439 bp, 348 bp) (Fig. 3)
SNP 3 (T/C located in Exon 37 loci 14420), revealed a non synonymous mutation, which caused change of amino acid, A1970G,
Alanine (GCC) to Valine (GUC). The observed heterozygosity of the
three genotyped SNPs with hardy Weinberg equilibrium P-value
and LD analysis are shown in Tables 6 and 7, respectively. Two
haplotypes were constructed using the three SNP information Haplotype 1 (C1A2T3) and Haplotype 2 (T1G2C3) excluding haplotypes
with frequencies <0.05 (Table 8).

3.4. Association analysis results

The results showed signicant association of SNP 1(911 C/T)
with myristic acid (p = 0.024), TT individuals had shown 0.83 wt%
(p = 0.003) and 0.60 wt% (P = 0.032) lower levels of myristic acid
than CT and CC individuals, respectively. Similarly it had shown
overall signicant (p = 0.025) associations with palmitic acid, with
TT individuals shown 1.33 wt% (p = 0.005) and 0.83 wt% (a tendency
of P = 0.08) lower levels of palmitic acid than CT and CC individuals,
respectively. A signicant (P = 0.05) association of SNP1 was also
observed with palmitoleic acid (C16:1). Furthermore TT individuals had shown 0.23 wt% higher (P = 0.005) levels of linoleic than CC
individuals.
An overall signicant effect (P = 0.44) of SNP 2(852 A/G) on fat
wt% was also observed in the present study. Meanwhile GG individuals showed 0.52 wt% (p = 0.049) and 0.96 wt% (P = 0.034) lower
levels of myristic acid and palmitic acid, respectively, than AG individuals. Moreover GG individuals showing 0.18 wt% higher linoleic
levels than TT individuals (p = 0.036).
SNP3 had CC individuals shown 0.96 wt% lower (p = 0.033)
palmitic acid than TC individuals. Furthermore an overall tendency
of signicance (p = 0.06) of SNP3 on linoleic acid (p = 0.038), with CC
individuals shown 1.43 wt% higher (a tendency of p = 0.07) levels
than TT individuals, was observed in the present study.
H1 (Haplotype 1) had shown a signicant (p = 0.049) effect
on linoleic acid, animals with 1 and 2 copies showing 0.14 wt%
(p = 0.021) and 0.12 wt% (p = 0.042), respectively, higher levels of
linoleic acid than animals with other copies.
Animals with one and two copies of haplotype 2 had shown
0.62 wt% (p = 0.032) and 0.51 wt% (p = 0.061), respectively, lower
levels of myristic acid than animals without the copy. Moreover,
animals with two copies of haplotype 2 had shown 0.97 wt% lower
(p = 0.05) levels of palmitic acid than animals with one copy of
the haplotype. Haplotype 2 also had shown signicant associations
with linoleic acid (p = 0.05), animals with two copies had shown to
contain 0.16 wt% higher (p = 0.043) linoleic acid levels than animals
without the copy.

4. Discussion
The high variability of fatty acids in space and time arise by
intrinsic (genetic and physiology) and extrinsic (nutrition and management) factors (Chilliard et al., 2003). Although limited by rumen
biohydrogenation, nutrition holds a predominant role in modifying milk fat composition. Nevertheless, medium to high heritability
of milk FAs confers a niche for genetic intervention (Bastin et al.,
2011). Hence, nutrition and genetics play their part in modulating
milk fat composition.

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

Fig. 1. PCR-RFLP results of Intron 1 911 C/T and reverse sequence (G/A) result of the SNP.

Fig. 2. PCR-RFLP results of Intron 2 852 A/G and reverse sequence (T/C) result.

Fig. 3. PCR-RFLP results of Exon 37 14420 T/C and forward sequence (T/C) result.

In our analysis the levels of the fatty acids were consistent to

results by Bernard et al. (2005), FA analysis on milk samples of
mid-lactation conducted in alpine goats fed sun ower oil. However, contrary to their result, in our study relatively higher levels
of oleic acids were observed, differences in stages of lactation (fat
mobilization of the early lactation in our study) and oil seed supplements between the two studies may have caused the observed
differences. Elevated levels of oleic acid is an apparent phenomenon
in early lactation, due to mobilization of FAs of adipose tissue to the
mammary gland (Pires et al., 2013). Low levels of short and medium
chain fatty acids (except C4:0 showed higher levels), low levels of
CLA and elevated levels of oleic acids were observed in our analysis
contrary to the study by Bouattour et al. (2008), on milk samples
collected 80 days postpartum (around pick lactation), in Saanen
goats fed soybean oil. Similarly the impact of lactation stage on body
fat mobilization and its consequential effects on de novo fatty acids
may have caused the observed differences. The high level of oleic
acid indicates high fat mobilization from the adipose tissue and possible higher oil supplement feeding. Due to physiological condition
and gene expression differences across lactation, fatty acid prole
varies widely throughout lactation (Atasoglu et al., 2009). In bovine
Stoop et al. (2009a) has also observed increasing levels of palmitic
(C16:0), stearic (C18:0) and oleic acids (C18:1 cis-9) in early lactation. Lactation is a high energy demanding process, hence the
increasing levels of those FAs in that particular time is due to their

mobilization from the body fat to mammary gland to replenish the

negative energy balance of the early lactation.
The observed low desaturation indices of C14:0 and C16:0 and
high desaturation and elongation indices of C18:0 indicate minimal
activity of delta 9 desaturation enzyme on saturated fatty acids
other than stearic acid and nonexistent elongation of C14:1 and
C16:1 into oleic acid. This indicates that there is no pathway of
modulating fatty acids through elongation of myristeloic acid into
oleic acid in early lactation. The same phenomenon was observed
in bovine (Nakov et al., 2013). The linolenic Omega-6 fatty acid
comprised around 3.18% of the fatty acids; this enriched level was
possibly due to oil seed feeding of the farms, whereas the linoleic
n-3 which comprised around 0.55%, predominantly C18:3 n-3, is
an indication of feeding vegetable oils. According to the Institute of
Medicine of the National Academy of Science in the United States
the recommended ratio of Omega-6 to Omega-3 for good health
was set below 10:1 (Allport, 2008), hence we can assume the ratio
in our study (7.21) ts the healthy standard. But the required ratio
also varies with the health and physiological state of the person
(Simopoulos, 2002).
In majority of the fatty acids herd and HTD had signicant effects
followed by parity, except in C4:0 which seemed to be affected
signicantly only by parity. This implies that feeding regimen and
management plays a big role in modulating the fatty acid prole of
dairy goats in early lactation (De La Fuente et al., 2009).

Table 3
Spearmans rho bivariate correlation of de novo with long chain fatty acids.

a
b

C17:1

C18:0

C18:1 trans-9

C18:1 cis-9

C18:2 n-6 trans

C18:2 n-6 cis

C20:0

C18:3 n-6

C18:3 n-3

C20:1 n-9

C18:2 cis-9
trans-11 CLA

C18:2 trans10,cis-12CLA

0.226b
0.140a
0.200b
0.269b
0.251b
0.300b
0.266b
0.276b
0.017
0.141a
0.088
0.347b
0.158b

0.154
0.017
0.513b
0.521b
0.583b
0.542b
0.560b
0.444b
0.650b
0.207b
0.417b
0.370b
0.574b

0.145
0.081
0.142a
0.283b
0.426b
0.583b
0.476b
0.527b
0.523b
0.772b
0.228b
0.474b
0.306b

0.004
0.014
0.253b
0.249b
0.221b
0.234b
0.177b
0.270b
0.077
0.124a
0.148a
0.073
0.052

0.047
0.177b
0.744b
0.798b
0.897b
0.784b
0.867b
0.738b
0.901b
0.325b
0.492b
0.413b
0.492b

0.165a
0.006
0.258b
0.205b
0.192b
0.131a
0.153b
0.112a
0.159b
0.230b
0.190b
0.087
0.192b

0.170a
0.030
0.374b
0.370b
0.365b
0.343b
0.320b
0.242b
0.325b
0.121a
0.000
0.204b
0.104

0.104
0.139a
0.238b
0.236b
0.268b
0.304b
0.269b
0.297b
0.197b
0.202b
0.040
0.034
0.088

0.094
0.121a
0.030
0.001
0.014
0.030
0.017
0.086
0.049
0.040
0.026
0.068
0.054

0.104
0.077
0.138a
0.122a
0.085
0.006
0.069
0.069
0.035
0.050
0.281b
0.028
0.042

0.196a
0.058
0.129a
0.147a
0.148a
0.143a
0.151b
0.132a
0.042
0.089
0.116a
0.039
0.009

0.035
0.155b
0.039
0.124a
0.162b
0.270b
0.170b
0.232b
0.096
0.292b
0.212b
0.029
0.109

0.062
0.041
0.102
0.147a
0.175b
0.233b
0.198b
0.332b
0.142a
0.244b
0.163b
0.040
0.068

Correlation is signicant at the 0.05 level (2-tailed).

Correlation is signicant at the 0.01 level (2-tailed).

Table 4
Spearmans rho bivariate correlation coefcient of long chain fatty acids.

C17:1
C18:0
C18:1 trans-9
C18:1 cis-9
C18:2 n-6 Trans
C18:2 n-6 cis
C20:0
C18:3 n-6
C18:3 n-3
C20:1 n-9
C18:2 cis-9 trans-11 CLA
C18:2 trans-10,cis-12 CLA
a
b

C17:0

C17:1

C18:0

C18:1 trans-9

C18:1 cis-9

C18:2 n-6 Trans

C18:2 n-6 cis

C20:0

C18:3 n-6

C18:3 n-3

C20:1 n-9

C18:2 cis-9 trans-11 CLA

0.100
0.262b
0.253b
0.082
0. 213b
0.013
0.373b
0.003
0.053
0.359b
0.439b
0.158b

0.143a
0.157b
0.726b
0.090
0.216b
0.172b
0.020
0.081
0.052
0.099
0.040

0.096
0.353b
0.164b
0.138a
0.315b
0.043
0.011
0.232b
0.367b
0.185b

0.191b
0.267b
0.080
0.093
0.332b
0.128a
0.081
0.368b
0.378b

0.231b
0.334b
0.234b
0.005
0.061
0.079
0.114
0.138a

0.303b
0.142a
0.056
0.266b
0.085
0.376b
0.161b

0.114a
0.044
0.341b
0.018
0.284b
0.105

0.029
0.049
0.562b
0.267b
0.082

0.047
0.068
0.040
0.305b

0.176b
0.100
0.046

0.227b
0.112

0.168b

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

FAT
C4:0
C6:0
C8:0
C10:0
C11:0
C12:0
C13:0
C14:0
C14:1
C15:0
C16:0
C16:1

C17:0

Correlation is signicant at the 0.05 level (2-tailed).

Correlation is signicant at the 0.01 level (2-tailed).

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

The observed high correlation among de novo fatty acids

(C6:0C14:0) indicates their similar source of origin. The distance
in carbon chain length within the group was equivalent with their
degree of correlation. The furthest the distance in their carbon chain
length the weaker the observed correlation. This is may be due
to the presence of less intervening factors between neighboring
fatty acids in the successive addition of acetyl group (two carbon
atoms) to the growing chain of fatty acids during de novo fatty acid
synthesis.
The analysis of correlation patterns have shown suppression of
the de novo fatty acids (C6:0C14:0) concomitant with rising levels of oleic acid. This result was supported by a study in goats fed
high oil seeds (Chilliard and Ferlay, 2004). The high content of oleic
acid, might have occurred due to high bioavailability of Carbon 18
fatty acids from high lipid supplement feeding and/or lipid mobilization of the early lactation (mainly 16:0, 18:0 and 18:1 cis-9)
(Chilliard et al., 2003). This phenomenon was paralleled with suppression of the de novo FA secretions, supposedly caused by the
inhibitory effects of the long chain fatty acids on ACACA, a rate limiting enzyme in de novo FA synthesis (Jacobs et al., 2013). Fatty
acids at the margins of the de novo fatty acid synthesis (C4:0 and
C16:0) were not seem to be affected by the phenomenon. Butyric
acid (C4:0) is a short chain de novo fat which has an anticarcinogenic effect in colon cancer (Parodi, 1999; Aluko, 2012). However, it
has showed negative correlations with its presumed source of origin, the de novo FA synthesis. And it showed positive correlation
with most of the MUFA and PUFA fatty acids, contrary to the correlation patterns of other de novo FAs. The inhibitory effect of long
chain FAs on de novo increased with the increase of carbon chain
length in the short and medium chain fatty acids (C4:0C14). This
showed that the effect of the suppression effect is more signicant
in the longer chain de novo fatty acids, as it is more demanding to
synthesize longer chains. Hence, the suppression of long chain fatty
acids on de novo may have a negligible (below threshold) effect on
C4:0, since its synthesis is less demanding. Possibly the unique correlation patterns of C4:0 may have arisen due to other metabolic
sources. Furthermore, the analysis of variance revealed that except
parity the other factors had no signicant effect on C4:0 suggesting
less control of extrinsic factors (C18 bio-availabilities either from
fat mobilization or diet) on its synthesis. This is further substantiated by the high heritability of C4:0 reported in bovine (Stoop et al.,
2008). Consistent levels of C4:0 while de novo fatty acids were suppressed was reported in studies conducted in bovine (Palmquist
et al., 1993; Parodi, 2006; Renn et al., 2013).
Palmitic acid has two sources of origin, the de novo fat synthesis
and performed fatty acids (adipose tissue and feed supplement). Its
had neither shown suppressed levels with the other de novo fatty
acids nor had the highest level as was reported in other studies.
Here, its dual source of origin is playing a role in masking the suppression of de novo fat synthesis by the ow of FAs from adipose
origin of early lactation. The observed weak negative correlations
with some de novo fatty acids (C6:0 and C8:0), indicates majority of
the palmitic acids were of adipose and dietary sources. This partly
explains why the levels of palmitic acid were lower than oleic acid.
The de novo side source of palmitic acid is suppressed; hence, its
quantity is not sufcient enough to exceed oleic acid. However, its
relatively higher quantity and its strong positive correlation with
C14:0 indicates de novo fatty still had a limited contribution to the
palmitic acid pool. Higher levels of oleic acid and high levels of
palmitic (in contrary to the suppression of de novo fat synthesis)
clearly explains high bioavailability of C16:0 and C18:1 cis-9 from
diet or fat mobilization of the early lactation. Furthermore, the elevated levels of oleic acid and decreasing levels of de novo fatty acids
can also be explained by the inhibitory effect of dietary fatty acids
or energy concentrates on rumen condition, which decreases the
pool of fermentable organic matter which might have shifted the

proportion of volatile fatty acids to high propionate/acetate, which

limit the availability of substrates (acetate and Beta hydroxybutyrate) to the mammary gland de novo fatty acid synthesis (Bauman
et al., 2004).The observed high desaturation index of C18:0 further indicated the potency of delta-9 desaturase activity, which is
capable of contributing more than 50% oleic acid (18:1, cis-9) by
desaturating stearic acid. This may also have contributed to the
elevated levels of oleic acid (Bickerstaffe et al., 1974).
Odd chain fatty acids (C11, C13, C15, and C17) have two
metabolic sources, the de novo fatty acid synthesis and rumen
bacteria (Vlaeminck et al., 2006; Dewhurst et al., 2007 Dewhurst
et al., 2007). The observed positive correlations with medium chain
fatty acids reected the partial same source of origin, the de novo
fatty acid synthesis (Vlaeminck et al., 2006). A good rumen condition favors the increase of rumen bacteria, with high amount
of them escaping the rumen, upon which digestion of their lipid
membrane supply an increased amount of odd chain fatty acids
(Vlaeminck et al., 2006). Meanwhile rumen condition is associated with the completeness of rumen biohydrogenation, and good
rumen condition gives a complete rumen biodyrogenation end
products, stearic acid and the trans fatty acids. Hence, OCFAs are
correlated to rumen biohydrogenation activity since both are the
consequential effects of rumen condition and occur concomitantly.
In our study, weve observed positive correlations of OCFAs with
CLA (through precursor vaccenic acd C18:1, trans-11), trans fatty
acids C18:1 trans-9, C18:2 n-6 trans, which are relatively end products of bioydrogenation. And negative correlations were observed
with their equivalent cis-biohydrogenation forms (C18:1 cis-9 and
C18:2 n-6 cis). This indicates associations of the odd chain fatty
acids with rumen biohydrogenation patterns. Hence, as rumen
bacterial population are reected in OCFAs, odd chain fatty acids
may indicate rumen function and patterns of rumen biohydrogenation (van Engelen, 2014). Negative correlations with oleic acid,
stearic acids and some cis-biohydrogenation intermediates were
also observed in dairy cattle (Vlaeminck et al., 2004). Interestingly,
all the OCFAs had shown strong positive correlations with rumenic
acid (CLA), presumably through precursor-product relationships
of vaccenic acid (a trans-biohydrogenation intermediate) (C18:1,
trans-11), which is a substrate for delta 9 desaturate enzyme for CLA
production. This further reinforces the associations of Odd chain
fatty acids with trans monoene intermediates. The negative correlation between odd chain fatty acids and stearic acids (end result of
rumen bio-hydrogenation) was observed in our study, this might
have happened due to the rate limiting, requiring high activation
energy, conversion of trans fatty acids to stearic acid, which might
have given trans-fatty acids a chance to accumulate and leave the
rumen (Shingeld et al., 2010). Moreover, there are two rumen
bacteria strains involving in hydrogenation, type A hydrogenating
from polyunsaturated fatty acids to trans fatty acids and type B from
trans to stearic acid (C18:0) (Harfoot and Hazlewood, 1997). Hence,
observed negative correlation of the odd chain fatty acids with
stearic acid might have happened, due to contradicting associations
of odd chain fatty acids with type B rumen bacteria.
The correlations of odd chain fatty acids with rumen biohydrogenation intermediates reinforced the supposition that OCFA
my serve as a diagnostic tool for rumen function and patterns of
biohydrogenation (Vlaeminck et al., 2006). From the patterns of
correlations we hypothesize that high bioavailability of polyunsaturated C18 fatty acids coupled with good rumen function, could
result in high levels of rumenic CLA content, which is implied in
some studies (Bouattour et al., 2008). This condition is more pronounced in pasture fed animals, due to their high PUFA content and
favoring rumen condition natures of pasture understory. The effect
of feeding grass silage, which facilitate efcient rumen condition,
was shown to increase milk OBCFA (odd and branched chain fatty
acids) and rumenic CLA (Patel et al., 2013).

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

Table 5
Inuence of factors on individual and groups of fatty acids.
SNP1
Fat Percentage
C4:0
C6:0
C8:0
C10:0
C11:0
C12:0
C13:0
C14:0
C14:1
C15:0
C16:0
C16:1
C17:0
C17:1
C18:0
C18:1 n9 trans
C18: 1 n9 cis
*
**
***

SNP2

SNP3

Herd

HTD

**

***

Parity
*
*

***
***

***

***

**

***

**

***

***

***

***

***

**

*
***

**

*
***

***

**

**

***
*

**

***

SNP1
C18:2 n6 cis
C20:0
C18:3 n3
C20:1 n9
cis-9,trans-11 CLA
trans-10,cis-12 CLA
MCFA
SFA
MUFA
PUFA
UFA
SFA/UFA
AI
Omega 3
Omega 6
Omega 6/Omega 3
GSF

SNP2

SNP3

Herd

HTD

Parity

***
*

***

***

***

***

**

*
***

***

***

**
***
***

***

***

***

***

***

***

***

***

***

***

***

**

***

P < 0.05.
P < 0.01.
P < 0.001.

Table 6
Marker summary.
#

Name

Position

Number of Geno.

ObsHETa

PredHETb

Alleles

MAFc

HWpvald

1
2
3

SNP1
SNP2
SNP3

911
852
14420

488
488
488

0.484
0.434
0.455

0.449
0.442
0.451

C:T
A:G
T:C

0.34
0.33
0.344

0.1115
0.7568
0.9669

a
b
c
d

Observed heterozygosity.
Predicted heterozygosity.
Minor allele frequency.
Hardy Weinberg p-value.

The degree of information of the markers used for inferences

was assured by their consistency with HWE, their high heterozygosity and linkage disequilibrium. Association analysis revealed,
a signicant association of FASN SNP1 and SNP2 particularly SNP
1(911 C/T, intron 1) with palmitic, myristic and linoleic fatty acids.
Linkage of myristic acid on BTA 19, where FASN is located was
reported in a study in bovine (Stoop et al., 2009b). Heterozygous
animals for the SNPs had shown higher myristic and palmitic acids,
and lower linoleic acids, followed by major homozygous genotypes.
The minor homozygous genotypes had shown higher health benets through the reduction of palmitic and myristic fatty acids and
increasing of linoleic acid. SNP 1 is located in intron 1 upstream
of transcription initiation, located in Exon 2. Hence, the SNP might
have an effect on the promoter of the gene; else it is in linkage
disequilibrium with other causal variants, while SNP 2 is located
in the intronic region two. SNP3, which is located in exon 37, had
shown only tendencies of associations, this explains SNPs located
in intronic region can be equally important in aiding genetic based
selections. Similarly H1 had shown higher levels of myristic and
palmitic fatty acids and lower levels of linoleic acids; to the contrary H2 had shown the opposite effects. Hence, haplotype 2 (TGC)
has a potential to improve milk fatty acid healthfulness. Despite
the amino acid changes caused by SNP3, it had shown weak association with the observed phenotypic traits as compared to the two
intronic SNPs, this suggest that introns could be equally important
in aiding marker assisted selection. Due to physiological and gene
expression differences across lactation, fatty acid prole varies
widely throughout lactation (Atasoglu et al., 2009). In mid and late
lactation, when negative energy balance dissipates, de novo fatty
acid has a strong contribution (Bionaz and Loor, 2008), hence, a
thorough analysis of the effects of the variants in different stages
of lactation will be decisive and meaningful.

Table 7
Linkage disequilibrium results.
Linkage Disequilibrium

D a

r2 b

SNP1-SNP2
SNP1-SNP3
SNP2-SNP3

90
86
91

77
72
78

a
b

D is correlation coefcient.
r2 is the statistical correlation.

Table 8
Haplotypes constructed with three intragenic SNPs.
Haplotypes

Genotype

Frequency

H1
H2

C1 A2 T3
T1 G2 C3

0.58%
0.36%

N.B. Only haplotypes greater or equal than 0.05% are considered in the analysis.

5. Conclusion
The FASN markers have shown signicant associations with
healthfulness of milk fatty acid. Our fatty acid analysis revealed
lower levels of de novo fatty acid synthesis concomitant with elevated levels of oleic acid. This indicates inhibitory effects of high
C18 fatty acids bioavailability on de novo fat synthesis in goat. Odd
chain fatty acids had shown distinct correlations with rumen biohydrogenation intermediates and CLA (supposedly through vaccenic
acid). This observed correlations could serve as potential diagnostic
tool for rumen condition and biohydrogenation patterns.
Acknowledgements
This research was jointly supported by a special fund, Agroscientic Research for the Public Interest (China; 201103038). The

10

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

authors wish to express appreciation to personnel of the Xinong

Saanen Dairy Goat farm and laboratory of the Xinong Saanen Dairy
Goat for their assistance.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.smallrumres.
2016.03.025.
References
Abbas, H.M., Hassan, F.A., El-Gawad, M.A.A., Enab, A., 2014. Physicochemical
characteristics of goats milk. Life Sci. J. 11.
Allport, S., 2008. The Queen of Fats: Why Omega-3s Were Removed from the
Western Diet and What We can Do to Replace Them. University of California
Press.
Aluko, R.E., 2012. Functional Foods and Nutraceuticals. Springer.
Atasoglu, C., Uysal-Pala, C., Karagl-Yceer, Y., 2009. Changes in milk fatty acid
composition of goats during lactation in a semi-intensive production system.
Arch. Anim. Breed. 52, 627636.
Barber, M.C., Clegg, R.A., Travers, M.T., Vernon, R.G., 1997. Lipid metabolism in the
lactating mammary gland. Biochim. Biophys. Acta Lipids Lipid Metab. 1347,
101126.
Barik, S., 2008. An intronic microRNA silences genes that are functionally
antagonistic to its host gene. Nucleic Acids Res. 36, 52325241.
Bastin, C., Gengler, N., Soyeurt, H., 2011. Phenotypic and genetic variability of
production traits and milk fatty acid contents across days in milk for Walloon
Holstein rst-parity cows. J. Dairy Sci. 94, 41524163.
Bauman, D.E., Pereld, J., Lock, A.L., 2004. Effect of trans fatty acids on milk fat and
their impact on human health. Proc. Southwest Nutr. Manage. Conf., 4152.
Belury, M.A., 2002. Inhibition of carcinogenesis by conjugated linoleic acid:
potential mechanisms of action. J. Nutr. 132, 29952998.
Berillo, O., Issabekova, A., Regnier, M., Ivashchenko, A.T., 2013. Characteristics of
Intronic and Intergenic Human miRNAs and Features of their Interaction with
mRNA. arXiv preprint arXiv:1301.5093.
Bernard, L., Rouel, J., Leroux, C., Ferlay, A., Faulconnier, Y., Legrand, P., Chilliard, Y.,
2005. Mammary lipid metabolism and milk fatty acid secretion in alpine goats
fed vegetable lipids. J. Dairy Sci. 88, 14781489.
Bickerstaffe, R., Annison, E., Linzell, J., 1974. The metabolism of glucose, acetate,
lipids and amino acids in lactating dairy cows. J. Agric. Sci. 82, 7185.
Bionaz, M., Loor, J.J., 2008. Gene networks driving bovine milk fat synthesis during
the lactation cycle. BMC Genomics 9, 366.
Bou, R., Codony, R., Tres, A., Decker, E.A., Guardiola, F., 2009. Dietary strategies to
improve nutritional value, oxidative stability, and sensory properties of
poultry products. Crit. Rev. Food Sci. Nutr. 49, 800822.
Bouattour, M., Casals, R., Albanell, E., Such, X., Caja, G., 2008. Feeding soybean oil to
dairy goats increases conjugated linoleic acid in milk. J. Dairy Sci. 91,
23992407.
Bouwman, A.C., Bovenhuis, H., Visker, M.H., van Arendonk, J.A., 2011.
Genome-wide association of milk fatty acids in Dutch dairy cattle. BMC Genet.
12, 43.
Chilliard, Y., Ferlay, A., 2004. Dietary lipids and forages interactions on cow and
goat milk fatty acid composition and sensory properties. Reprod. Nutr. Dev. 44,
467492.
Chilliard, Y., Ferlay, A., Rouel, J., Lamberet, G., 2003. A review of nutritional and
physiological factors affecting goat milk lipid synthesis and lipolysis. J. Dairy
Sci. 86, 17511770.
Chouinard, P.Y., Corneau, L., Barbano, D.M., Metzger, L.E., Bauman, D.E., 1999.
Conjugated linoleic acids alter milk fatty acid composition and inhibit milk fat
secretion in dairy cows. J. Nutr. 129, 15791584.
De La Fuente, L., Barbosa, E., Carriedo, J., Gonzalo, C., Arenas, R., Fresno, J., San
Primitivo, F., 2009. Factors inuencing variation of fatty acid content in ovine
milk. J. Dairy Sci. 92, 37913799.
Decker, E.A., Park, Y., 2010. Healthier meat products as functional foods. Meat Sci.
86, 4955.
Dekkers, J.C., 2004. Commercial application of marker-and gene-assisted selection
in livestock: strategies and lessons. J. Anim. Sci. 82, E313E328.
Dewhurst, R., Moorby, J., Vlaeminck, B., Fievez, V., 2007. Apparent recovery of
duodenal odd-and branched-chain fatty acids in milk of dairy cows. J. Dairy Sci.
90, 17751780.
Du, F.-X., Clutter, A.C., Lohuis, M.M., 2007. Characterizing linkage disequilibrium in
pig populations. Int. J. Biol. Sci. 3, 166.
Enig, M.G., 2003. Milk homogenization and heart disease. Wise Traditions in Food,
Farming, and the Healing Arts (accessed 09.01.14.) http://www.realmilk.com/
health/milk-homogenization-and-heart-disease/.
Hara, A., Radin, N.S., 1978. Lipid extraction of tissues with a low-toxicity solvent.
Anal. Biochem. 90, 420426.
Harfoot, C., Hazlewood, G., 1997. Lipid metabolism in the rumen. In: The Rumen
Microbial Ecosystem. Springer, pp. 382426.
Hayes, B., Goddard, M., 2001. Prediction of total genetic value using genome-wide
dense marker maps. Genetics 157, 18191829.

Hillgartner, F.B., Salati, L.M., Goodridge, A.G., 1995. Physiological and molecular
mechanisms involved in nutritional regulation of fatty acid synthesis. Physiol.
Rev. 75, 4776.
Horie, T., Ono, K., Nishi, H., Iwanaga, Y., Nagao, K., Kinoshita, M., Kuwabara, Y.,
Takanabe, R., Hasegawa, K., Kita, T., 2009. MicroRNA-133 regulates the
expression of GLUT4 by targeting KLF15 and is involved in metabolic control in
cardiac myocytes. Biochem. Biophys. Res. Commun. 389, 315320.
Hui, Y.H., Chandan, R., 2007. Handbook of Food Products Manufacturing. Wiley.
Jacobs, A., Dijkstra, J., Liesman, J., VandeHaar, M., Lock, A., van Vuuren, A., Hendriks,
W., van Baal, J., 2013. Effects of short-and long-chain fatty acids on the
expression of stearoyl-CoA desaturase and other lipogenic genes in bovine
mammary epithelial cells. Animal 7, 15081516.
Jung, S., Choe, J.H., Kim, B., Yun, H., Kruk, Z.A., Jo, C., 2010. Effect of dietary mixture
of gallic acid and linoleic acid on antioxidative potential and quality of breast
meat from broilers. Meat Sci. 86, 520526.
Kim, J., Hwangbo, J., Choi, N.-J., Park, H., Yoon, D.-H., Park, E.-W., Lee, S.-H., Park,
B.-K., Kim, Y., 2007. Effect of dietary supplementation with conjugated linoleic
acid, with oleic, linoleic, or linolenic acid, on egg quality characteristics and fat
accumulation in the egg yolk. Poult. Sci. 86, 11801186.
Li, C., Sun, D., Zhang, S., Wang, S., Wu, X., Zhang, Q., Liu, L., Li, Y., Qiao, L., 2014.
Genome wide association study identies 20 novel promising genes associated
with milk fatty acid traits in Chinese Holstein. PLoS One 9, e96186.
Luna, P., Jurez, M., De la Fuente, M., 2005. Validation of a rapid milk fat separation
method to determine the fatty acid prole by gas chromatography. J. Dairy Sci.
88, 33773381.
Luo, Z., Wu, C.-I., 2001. Modeling linkage disequilibrium between a polymorphic
marker locus and a locus affecting complex dichotomous traits in natural
populations. Genetics 158, 17851800.
Matsuhashi, T., Maruyama, S., Uemoto, Y., Kobayashi, N., Mannen, H., Abe, T.,
Sakaguchi, S., Kobayashi, E., 2010. Effects of FASN, SCD, SREBP1 and GH gene
polymorphisms on fatty acid composition and carcass traits in Japanese Black
cattle. J. Anim. Sci., 20103121.
Matsumoto, H., Inada, S., Kobayashi, E., Abe, T., Hasebe, H., Sasazaki, S., Oyama, K.,
Mannen, H., 2012. Identication of SNPs in the FASN gene and their effect on
fatty acid milk composition in Holstein cattle. Livest. Sci. 144, 281284.
McNamara, J.P., Hillers, J.K., 1986. Regulation of bovine adipose tissue metabolism
during lactation. 2. Lipolysis response to milk production and energy intake. J.
Dairy Sci. 69, 30423050.
Morris, C.A., Cullen, N.G., Glass, B.C., Hyndman, D.L., Manley, T.R., Hickey, S.M.,
McEwan, J.C., Pitchford, W.S., Bottema, C.D., Lee, M.A., 2007. Fatty acid
synthase effects on bovine adipose fat and milk fat. Mamm. Genome 18, 6474.
Murphy, M., Uden, P., Palmquist, D., Wiktorsson, H., 1987. Rumen and total diet
digestibilities in lactating cows fed diets containing full-fat rapeseed. J. Dairy
Sci. 70, 15721582.
Nakov, R., 2010. Genetic regulation of bovine milk fatty acid composition:
Improving the healthfulness of milk through selection.
Nakov, R., Schoonmaker, J., Korn, K., Noack, K., Garrick, D., Koehler, K.,
Minick-Bormann, J., Reecy, J., Spurlock, D., Beitz, D., 2013. Sterol regulatory
element binding transcription factor 1 (SREBF1) polymorphism and milk fatty
acid composition. J. Dairy Sci. 96, 26052616.
Nakaya, A., Isobe, S.N., 2012. Will genomic selection be a practical method for plant
breeding? Ann. Bot. (mcs109).
Ordovas, L., Roy, R., Pampn, S., Zaragoza, P., Osta, R., Rodriguez-Rey, J.C., Rodellar,
C., 2008. The g. 763G > C SNP of the bovine FASN gene affects its promoter
activity via Sp-mediated regulation: implications for the bovine lactating
mammary gland. Physiol. Genomics 34, 144148.
Palmquist, D., 2006. Milk Fat: Origin of Fatty Acids and Inuence of Nutritional
Factors Thereon. Advanced Dairy Chemistry Volume 2 Lipids. Springer, pp.
4392.
Palmquist, D., Weisbjerg, M.R., Hvelplund, T., 1993. Ruminal, intestinal, and total
digestibilities of nutrients in cows fed diets high in fat and undegradable
protein. J. Dairy Sci. 76, 13531364.
Parodi, P., 1999. Conjugated linoleic acid and other anticarcinogenic agents of
bovine milk fat. J. Dairy Sci. 82, 13391349.
Parodi, P., 2006. Nutritional Signicance of Milk Lipids. Advanced Dairy Chemistry
Volume 2 Lipids. Springer, pp. 601639.
Patel, M., Wredle, E., Bertilsson, J., 2013. Effect of dietary proportion of grass silage
on milk fat with emphasis on odd-and branched-chain fatty acids in dairy
cows. J. Dairy Sci. 96, 390397.
Pires, J., Delavaud, C., Faulconnier, Y., Pomies, D., Chilliard, Y., 2013. Effects of body
condition score at calving on indicators of fat and protein mobilization of
periparturient Holstein-Friesian cows. J. Dairy Sci. 96, 64236439.
Raadsma, H.W., 2004. Quantitative trait loci mapping in dairy cattle: review and
meta-analysis. Genet. Sel. Evol. 36, 163190.
Renn, F.P., Freitas Jnior, J.E.d., Gandra, J.R., Verdurico, L.C., Santos, M.V.d.,
Barletta, R.V., Venturelli, B.C., Vilela, F.G., 2013. Fatty acid prole and
composition of milk protein fraction in dairy cows fed long-chain unsaturated
fatty acids during the transition period. Rev. Bras. Zootecnia 42, 813823.
Rhode, C., 2013. Signatures of Selection in Natural and Cultured Abalone (Haliotis
Midae): A Population Genomics Study. Stellenbosch University, Stellenbosch.
Sanz, A., Serrano, C., Calvo, J.H., Zaragoza, P., Altarriba, J., Rodellar, C., 2013. A single
nucleotide polymorphisms in the 5 UTR ovine FASN gene is associated with
milk fat yield. In: 64th Annual Meeting of the European Federation of Animal
Science. Book of Abstracts of the 64th Annual Meeting of the European
Federation of Animal Science, 19, 2013, 315. Nantes (France).

A.B. Haile et al. / Small Ruminant Research 138 (2016) 111

Sanz Ceballos, L., Sanz Sampelayo, M., Gil Extremera, F., Rodriguez Osorio, M.,
2009. Evaluation of the allergenicity of goat milk, cow milk, and their lactosera
in a guinea pig model. J. Dairy Sci. 92, 837846.
Schennink, A., Bovenhuis, H., Lon-Kloosterziel, K., Van Arendonk, J., Visker, M.,
2009. Effect of polymorphisms in the FASN, OLR1, PPARGC1A, PRL and STAT5A
genes on bovine milkfat composition. Anim. Genet. 40, 909916.
Shingeld, K.J., Bernard, L., Leroux, C., Chilliard, Y., 2010. Role of trans fatty acids in
the nutritional regulation of mammary lipogenesis in ruminants. Animal 4,
11401166.
Silanikove, N., Leitner, G., Merin, U., Prosser, C., 2010. Recent advances in
exploiting goats milk: quality, safety and production aspects. Small Ruminant
Res. 89, 110124.
Simopoulos, A.P., 2002. The importance of the ratio of omega-6/omega-3 essential
fatty acids. Biomed. Pharmacother. 56, 365379.
Stoop, W., Van Arendonk, J., Heck, J., Van Valenberg, H., Bovenhuis, H., 2008.
Genetic parameters for major milk fatty acids and milk production traits of
Dutch Holstein-Friesians. J. Dairy Sci. 91, 385394.
Stoop, W., Bovenhuis, H., Heck, J., Van Arendonk, J., 2009a. Effect of lactation stage
and energy status on milk fat composition of Holstein-Friesian cows. J. Dairy
Sci. 92, 14691478.
Stoop, W., Schennink, A., Visker, M., Mullaart, E., Van Arendonk, J., Bovenhuis, H.,
2009b. Genome-wide scan for bovine milk-fat composition. I. Quantitative
trait loci for short-and medium-chain fatty acids. J. Dairy Sci. 92, 46644675.
Tyagi, A., Kathirvelan, C., 2006. Conjugated linoleic acid in milk and its
anticarcinogenic potency. Indian Dairyman 58, 25.

11

Ulbricht, T., Southgate, D., 1991. Coronary heart disease: seven dietary factors.
Lancet 338, 985992.
van Engelen, S., 2014. Genome wide association studies for milk fatty acids as a
basis for methane prediction. 10th World Congress on Genetics Applied to
Livestock Production.
Vlaeminck, B., Fievez, V., Van Laar, H., Demeyer, D., 2004. Prediction of rumen
volatile fatty acid proportions produced in vitro using variations in rumen odd
and branched chain fatty acids. J. Anim. Physiol. Anim. Nutr. 88, 401411.
Vlaeminck, B., Fievez, V., Tamminga, S., Dewhurst, R., Van Vuuren, A., De
Brabander, D., Demeyer, D., 2006. Milk odd-and branched-chain fatty acids in
relation to the rumen fermentation pattern. J. Dairy Sci. 89, 39543964.
Von Eckardstein, A., 2006. Atherosclerosis: Diet and Drugs. Springer.
Whigham, L.D., Watras, A.C., Schoeller, D.A., 2007. Efcacy of conjugated linoleic
acid for reducing fat mass: a meta-analysis in humans. Am. J. Clin. Nutr. 85,
12031211.
Yun, Y., Adesanya, T.A., Mitra, R.D., 2012. A systematic study of gene expression
variation at single-nucleotide resolution reveals widespread regulatory roles
for uAUGs. Genome Res. 22 (June), 10891097, http://dx.doi.org/10.1101/gr.
117366.110, Published in advance March 27, 2012.
Zhu, J., Luo, J., Wang, W., Yu, K., Wang, H., Shi, H., Sun, Y., Lin, X., Li, J., 2014.
Inhibition of FASN reduces the synthesis of medium-chain fatty acids in goat
mammary gland. Animal 8, 14691478.