Sex and tissue specific DNA methylationpatterns in the house sparrow (Passerdomesticus)Yuming Shi

(1)

Sex and tissue specific DNA methylation patterns in the house sparrow (Passer domesticus)

Yuming Shi

Degree project inbiology, Bachelor ofscience, 2021 Examensarbete ibiologi 15 hp tillkandidatexamen, 2021

Biology Education Centre and Department ofEcology and Genetics, Uppsala University Supervisor: Arild Husby

(2)

Abstract

DNA methylation patterns are sex and tissue specific in many species, yet many studies use blood samples, due to its accessibility, to establish links between the DNA methylation and different phenotypes. This raises the question of whether DNA methylation in blood samples reflect the DNA methylation pattern in other tissues that are more relevant to the phenotype being studied. In this research, samples were collected from the brain, blood, liver and gonad of 16 house sparrow (Passer domesticus), half of them were female, while the others were male. Reduced representation bisulfite sequencing (RRBS) was performed to get the

methylation profile in each sample. The result showed a tissue specific methylation profile in the four investigated tissues, a strong and positive correlation between 0.74 – 0.85 was found between tissues, in which a weaker correlation was found between blood and other tissue. In differential methylation analysis, most of the differently methylated sites between sexes were found in gonads, while the fewest was found in blood, and Z chromosome was

overrepresented place in all four tissues where the majority of the differently methylated sites between sexes were found. Comparison with the house sparrow genome annotation found about half of the differentially methylated sites between sexes were within genes and about 20 % of them were in the exon or coding region of a gene. The result suggested that blood could be useful in reflecting the general DNA methylation level in other tissues, but it was not a reliable bioindicator for further detailed study in DNA methylation pattern or in gene

ontology enrichment pathway analysis.

Keywords: DNA methylation, tissue-specific pattern, sex-specific pattern, house sparrow

(3)

2

Epigenetic modifications, as a kind of modification on genome controlling the phenotype of the organism without altering the DNA sequence, have long been recognized as one of the most important ways to control gene expression in organisms (Suzuki and Bird 2008). There is a variety of epigenetic modifications, including DNA methylation, histone modification and chromatin structure. Most research has focused on DNA methylation, partly because of the ease at which DNA methylation information can be obtained and partly because of its ubiquitous presence in many species (Kilvitis et al. 2019), which I will also focus on here. In vertebrates, methylation mostly occurs in places where a cytosine is followed by a guanine, the so-called CpG sites (Deaton and Bird 2011). DNA methylation has divergent function in organisms. In plants, methylation can both silence and promote gene expression. For example, the Lcyc gene in the mutant toadflax (Linaria vulgaris) is extensively methylated and transcriptionally silent, causing a morphological difference in the flowers of the mutant (Cubas, Vincent, and Coen 1999). In Arabidopsis when DNA methylation is in the DNA methylation monitoring sequence (MEMS) within the promoter region of the gene ROS1, it can promote the expression of ROS1 (Xiao et al. 2019). In reptiles like turtle, DNA methylation can take part in temperature- dependent sex determination by suppressing the expression of some hormone pathway in the embryo (Radhakrishnan et al. 2018). In general, when methylation takes place in CpG sites near the transcription start site of genes, it usually plays a role of suppressor in gene expression regulation in many vertebrates, including in birds (Laine et al. 2016).

(5)

4

Recent studies have revealed that DNA methylation is one of the most important mechanisms regulating gene expression in response to environmental changes (Leenen, Muller, and Turner 2016). DNA methylation can help organisms quickly adapt to environmental changes and stimuli within their lifetime (Herrera, Pozo, and Bazaga 2012). Therefore, DNA methylation can play an important role in evolution, by providing a potential mechanism for phenotypic plasticity or developmental flexibility (Schlichting and Wund 2014). Thus, evidence regarding the significance of DNA methylation in ecology and evolution has been reported across a wide range of groups including bats (Liu et al. 2012), fish (Massicotte, Whitelaw, and Angers 2011), fungi (Herrera, Pozo, and Bazaga 2012) and plants (Schrey et al. 2013).

1.2 Measuring DNA methylation

In the early studies, DNA methylation was measured using a PCR-based method called methylation sensitive amplified fragment length polymorphism (MS-AFLP), where two restriction enzymes, the methylation insensitive MSPI and the sensitive HPAII, were applied (Liebl et al. 2013). Though efficient and inexpensive, this method can only provide us with genome-wide epigenetic variation rather than more information regarding the exact position or the function of these differences (Schrey et al. 2013). Therefore, it is not an ideal method when we try to conduct downstream functional analysis.

Nowadays, another widely used method in detecting DNA methylation is bisulfite sequencing (Suzuki and Bird 2008), which provides more specific quantitative details about DNA

(6)

5

methylation, such as the location of those variant sites and their nearby sequences. Bisulfite can convert unmethylated cytosine into uracil, thus, by comparing the sequencing results before and after bisulfite conversion, we can then identify the exact position of the methylated sites in the genome. There are two kinds of commonly used bisulfite sequencing assays, whole genome bisulfite sequencing (WGBS) and reduced representation bisulfite sequencing (RRBS). RRBS is more cost efficient and enriched for CG islands across genome compared to WGBS (Yong, Hsu, and Chen 2016). It is important to note that also bisulfite sequencing has some disadvantages, mostly notably that it causes damage to the DNA sequences, resulting a higher level of fragmentation, which might affect downstream analysis (Yong, Hsu, and Chen 2016).

To solve this problem, more and more new methods has been developed. One of the most intriguing methods is called ten-eleven translocation (TET)-assisted pyridine borane sequencing (Liu et al. 2019). This bisulfite-free method can preserve DNA fragments over 10 kilobases long, with a more even coverage over the genome and lower sequencing costs (Liu et al. 2019). Electrochemical impedance spectroscopy aided by monoclonal antibody is another new bisulfite-free methylation measuring measurement (Schiefelbein et al. 2018). This method is based on enzyme‐linked immunosorbent assay (ELISA) and is efficient enough, but its ability in determining the position of methylated cytosine might not be good enough (Schiefelbein et al. 2018).

(7)

6

1.3 Tissue specific DNA methylation patterns

While DNA methylation is frequently studied in vertebrate non-model organisms these days the vast majority of samples use blood (Husby 2020). However, it is well known that DNA methylation patterns differ between tissues due to their roles in cell differentiation (Bock et al.

2012) and in organ formation (Elder and McGraw 2020). In human, the pattern across cell and tissues had been proved to be different, and different cells within the same tissue could show a more similar pattern in methylation (Kundaje et al. 2015). While in birds, such as chicken (Gallus gallus), the gene controlling DNA methylation was found to have a tissue specific expression pattern in the hepatic glucocorticoid receptors gene, which has an essential role in regulating the nutritional stress response, such tissue specific difference could lead to a variant response to acute and chronic stress (Kang, Madkour, and Kuenzel 2017). A study focused on European starlings (Sturnus vulgaris) found that for the promoter region of Nr3c1, a glucocorticoid receptor gene, methylation pattern in blood and two brain regions, hippocampus as well as hypothalamus, varied significantly (Siller and Rubenstein 2019). The methylation level in blood was the lowest and the cluster based on the methylation profile of blood did not show a correlation with the other two tissue (Siller and Rubenstein 2019). As a result, it is often unclear to what extent any observed association between blood DNA methylation pattern and a phenotype is reflected in other tissues and also is the same in the two sexes.

1.4 Aims of this study

The aim of this work was to provide insights into tissue and sex specific patterns of DNA

(8)

7

methylation using Reduced Representation Bisulfite Sequencing (RRBS) dataset of different tissues from male and female house sparrow (Passer domesticus). Since most studies use blood samples to investigate the relationship between DNA methylation and phenotype, it is important to provide information about whether blood samples are representative for DNA methylation patterns. For example, if blood DNA methylation is very different from that seen in other tissues, it would bring into question results that have used DNA methylation in blood samples when studying phenotypes that are not directly related to blood. Such tissue differences can also be sex-specific, further complicating potential inferences and will also be examined here. My hypothesis was that the DNA methylation pattern in blood is different from those in other tissues of house sparrow and it cannot correctly reflect the DNA methylation profile in other tissues.

(9)

8

2. Material and Methods

2.1 Study system

A total of 16 house sparrows (8 female and 8 male) aged from 1 to 4 years old were captured in nets on February 16^th, 2014 in at Linesøya and Stokkøya on the coast outside Trondheim, Norway. At capture, their sexes and phenotypes (body mass, tarsus length, bill depth and length, wing length as well as bib size in males) were recorded and sparrows were ringed. Following terminal sampling, four different tissues were collected: brains, livers, blood, and gonads were collected and immediately snap frozen at -80 ºC to preserve RNA for future analysis. As the birds were collected in late winter, they were likely to be in non-breeding stage.

2.2 Method

2.2.1 Sample processing

About 20 mg tissue section was excised from each frozen tissue sample for genomic DNA extraction, after which the sample was immediately returned to -80 ºC storage to preserve RNA for later analysis. Genomic DNA was extracted using the Qiagen DNeasy Blood and Tissue Kit (Qiagen, USA) following manufacturer’s instructions. Each tissue sample was lysed in proteinase K for 7 hours at 56 ºC on a rocking platform. To obtain RNA free genomic DNA, RNase treatment was performed by adding 12 μl of 10 mg/ml RNaseA (ThermoFisher Scientific, USA) to each sample lysate and incubating for 20 minutes at room temperature before proceeding with extraction. DNA was eluted in 100 μl of AE Buffer. DNA was quantified using

(10)

9

the Qubit dsDNA HS Assay Kit (Invitrogen, USA) to ensure a minimum concentration of 10 ng/μl before sending to SciLife (Uppsala University, Sweden) for RRBS sequencing.

2.2.2 Sequencing and bioinformatic processing

To acquire DNA methylation, reduce representation bisulfite sequencing (RRBS) was performed. Library preparation and sequencing were performed at the SciLife Lab (Uppsala University, Sweden). To produce RRBS libraries, the manufacturer’s preparation protocol (Illumina) was used (Boyle et al. 2012). Samples were digested using the restriction enzyme MspI. Then, the products, DNA fragments of multiple size, were subsequently treated with bisulfite, converting un-methylated cytosine bases into uracil bases, while methylated cytosine bases would not be affected. DNA fragments were then end-repaired with DNA polymerase I, while for adapter ligation, A-overhangs were added to the 3′ ends of each fragment. Individual sample libraries were marked using standard Illumina adapters barcodes. Libraries were then purified, selected by size with Ampure XP beads (Beckman Coulter) and the concentration was determined by quantitative polymerase chain reaction (qPCR). The filtration produced fragment size range of approximately 30–180 base pairs (bp), with a mean of 85 bp. Into the same sequencing lane, six libraries were pooled. Using a HiSeq2500 sequencer with a HiSeq SBS sequencing kit version 4 (Illumina), each pool was sequenced 100 bp single end. Sequencing was performed in two separate runs to yield enough coverage for each sample. To obtain reliable sequence generation in the sequencing processing, the genome of PhiX was used as an internal positive control. Before data analysis, the reads and adapters from PhiX were removed from the dataset.

(11)

10

The quality of the sequencing reads was investigated with FastQC 0.11.7. (Andrew 2018). Low quality bases as well as Illumina adapter contamination resulting from readthrough of short fragments were trimmed using Trim Galore! v0.4.4 (Krueger 2017) with default parameters under the rrbs mode. Trimmed sequencing reads were aligned against the house sparrow reference genome v1.0 (NCBI accession number GCA_001700915.1, https://www.ncbi.nlm.nih.gov/genome/17653?genome_assembly_id=281679) using Bismark 0.19.1 aligner (Krueger and Andrews 2011) in rrbs mode.

2.2.3 Statistical data analysis

R package methylKit (Akalin et al. 2012) and cocor (Diedenhofen and Musch 2015) was used in R version 4.0.4 (R Core Team 2020) for statistical analysis, the script can be found in Appendix 1. The methylation level for every CpG site was calculated and compared in each tissue and sex group. CpG sites with 10x coverage were included for downstream differential methylation analysis (Leenen, Muller, and Turner 2016). All the data were assigned into groups as their sample parts, which is brain, blood, liver, and gonad (data from testicles and ovaries).

To test the similarity of methylation profile in all the samples, a cluster and a Principal component analysis (PCA) were conducted using functions in methylKit (Akalin et al. 2012).

The sample clustering analysis was performed using the ward method provided in methylKit.

Distance in the clustering plot represented the Pearson’s correlation coefficient (r) of methylation pattern between two samples. PCA was conducted to measure how all the

(12)

11

components affecting the variance in methylation pattern. To test whether the correlation coefficients were significantly different, the R package cocor (Diedenhofen and Musch 2015) was used. Besides, the percentage of differently methylated sites between sexes on every chromosome were calculated. The logistic regression model (equation 1 below) was used to calculate p-values in methylKit.

log 𝜋_𝑖

1 − 𝜋_𝑖 = 𝛽₀+ 𝛽₁𝑇𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡 _𝑖 (1)

πi is the methylation proportion of a CpG is, and the logistic regression would model the log odds ratio using the “Treatment” variable indicting which group the CpG regions belongs to.

The p-value was later adjusted to q-value using SLIM method (Wang, Tuominen, and Tsai 2011). When the difference in methylation of a site was larger than 15 % between sexes and the q-value < 0.01, it was considered as a significantly differentially methylated site (Lindner et al. 2021).

2.2.4 Gene Annotation

Differentially methylated CpG sites between tissues or sexes were annotated using BEDtools (Quinlan and Hall 2010), the used command could be found in Appendix 2. A list of significantly different site in methylation pattern were retrieved from the differently methylated region calculation above in 2.2.3. The list was compared with the house sparrow reference genome annotation, any site overlapped with the reference genome by at least 1 bp would be returned and annotated. Statistical tests was then performed in OriginLab (OriginLab Corporation 2017).

(13)

12

3. Result

3.1 Tissue and sex specific methylation pattern

As shown in Fig.1, samples from homologous tissues, i.e., samples from the same kind of organs in different individuals, had a higher similarity in methylation pattern. Sample names were given at the bottom of the figure, and they all ended with one or two letter(s) as the labels indicating where they were sampled from. Five clustering groups could be observed in Fig.1 based on their similarity. Among which, samples from ovaries and testicles were more similar to each other than others. Methylation profile in the blood sample of house sparrows were the most different ones when comparing with the profile in gonads, while patterns in livers were the most similar. A sex specific methylation pattern was also evident in Fig.1. The sample names in blue represented the samples from males, while red ones stood were samples females. Significant methylation pattern

difference between two sexes could be found in gonads, while in other tissues, no clear difference between sexes could be observed.

blood brain liver testicle ovary

(14)

13

Fig.1 Sample cluster based on methylation profile (BL = blood, B = brain, L = liver, T = testicle, O = ovary).

The correlation coefficients (r) between tissues were strong and positive as shown in Table.1.

The average r between samples from different tissues ranged from 0.74 to 0.85, and the r between samples from the same tissue ranged from 0.91 to 0.93. The correlation between blood and other tissues, especially between blood and liver, was weaker than other tissues.

Table 1. Mean correlation coefficient in methylation level between and within tissues Brain Blood Liver Testicle Ovary

Brain 0.93 0.79 0.82 0.79 0.85

Blood 0.79 0.91 0.74 0.80 0.81

Liver 0.82 0.74 0.92 0.82 0.85 Testicle 0.79 0.80 0.82 0.91 0.84

Ovary 0.85 0.81 0.85 0.84 0.91

The result of principal component analysis (PCA) was shown in Fig.2. A total of 64 components affecting the variance in DNA methylation patterns were calculated. Among which, the first principal component (PC1) accounted for 17% of the total variance and can largely separate brain samples from other tissues. The second principal component (PC2) accounted for 14% of the total variance and separated blood and brain samples from other tissues. Other principal components explained between 3 % to 10% of the variance. As displayed in Fig.2, methylation patterns in house sparrow livers were more similar to those in their gonads, especially to ovaries in females. Plus, the distribution of samples from brain suggested that the relatively larger variance between them than from samples in other tissues.

(15)

14

Fig.2 PCA analysis on methylation pattern difference in all samples

3.2 Differently methylated regions and their function

In Table 2, the methylation differences between sexes in four tissues is shown. As

described in 2.3, significantly different (q < 0.01) sites were those sites with at least 15 % difference between male and female. Females were used as the “control” here, therefore, hypermethylated stood for cases where males had a significantly higher DNA methylation

at a site, while hypomethylated meant that DNA methylation in females were higher.

Significant sex difference in the number of CpG sites that were hyper- or hypomethylated was found in all tissues except in liver (Table 2), suggesting a different tissue-specific pattern in hypermethylation and hypomethylation. The highest sex difference was, not unexpectedly, found in gonads, while the lowest was in blood. Males had a higher number of DNA hypermethylated sites in brain and blood, while females have more in gonads. In

testicles

brain

ovary and liver

blood

(17%) (14%)

(16)

15

liver, the was no significant sex difference.

Table 2. The level of hypermethylation (male > female) and hypomethylation (male < female) in four tissues

Tissue Hypermethylated Hypomethylated χ² p-value

Number of sites

Percentage (%)

Number of sites

Percentage (%)

brain 982 0.355 748 0.270 31.651 < 0.01

blood 423 0.230 273 0.148 32.328 < 0.01

liver 972 0.234 900 0.235 2.769 0.096

gonad 7454 4.163 11491 6.418 860.25 < 0.01

all 9831 13412 384.4 < 0.01

In Fig.3, the distribution of the differently methylated regions between sexes over the chromosomes of house sparrow tissue samples is displayed. On the y-axis, the sequence names of chromosomes of house sparrow in GenBank database were shown, and the percentage of significantly different sites (q < 0.01) with at least 15% difference between male and female) over a chromosome was displayed on the x-axis. In general, I found a considerably higher genome-wide methylation difference between sexes in gonad as expected (notice that the scale in gonad’s figure is much larger than the others), while the level in blood was much lower than the level in other tissues, as illustrated in Fig.3 and Table 1. In addition, the Z chromosome (CM004527.1, the longest bar in all four graphs) was a common place where most methylation pattern difference was observed in all four kinds of tissues, as suggested in Fig.3.

(17)

16

Fig.3 Chromosomes with at least 15% difference in methylated percentage, hypermethylated refers to site where males had a significantly higher DNA methylation level, hypomethylated means females had higher.

After comparison with the house sparrow genome annotation, 5971 matches were discovered out of a total of 9831 divergent methylated sites from the four investigated tissues. Among all the matches, 3655 (61.2 %) were from gonad samples, 1114 (18.6 %) were from livers, 970 (16.2 %) came from brains, and 232 (4 %) were from blood. About 14 – 20 % of the

differently methylated sites were found in the coding sequence (CDS) of a mRNA and exon of a gene, and approximately 29 – 34 % of the sites were from the mRNA and gene region on the genome. Less than 1 % of the differently methylated sites were in the untranslated region (UTR) on the 3’ and 5’ of a mRNA. It is worth mentioning that we currently do not know much about these UTRs, therefore, many of them were not annotated in the reference genome.

blood

gonad brain

liver

(18)

17

Fig.4 Functional annotation of the differentially methylated region in four tissues.

brain blood

liver gonad

(19)

18

4. Discussion

The main findings in this work were as follows. Firstly, a tissue-specific DNA methylation pattern existed in the house sparrow, higher similarity can be found in homologous tissues, like between testicles and ovaries. As illustrated in Fig.1, the profile in blood was more similar to the profile in brain than those in liver and gonad, while gonads had a more similar DNA methylation profile to that in liver. In an epigenomics research done on human, the difference in DNA methylation level was found to be correlated with the cell origin, cells could share a more similar DNA methylation profile with cells in other tissues if they had the same origin (Kundaje et al. 2015). Therefore, we could infer that gonad and liver might have a similar origin in the embryo of house sparrow, given their similarity in DNA methylation pattern. In previous research, cells in liver were found to have a mixed origin, from both endoderm and mesoderm (Zaret 2008). While, cells in gonads were discovered to have a primary mesoderm origin (Satoh 1991). These cells derived from mesoderm may contribute to the similarity in DNA methylation pattern between liver and gonad of house sparrow. And as shown in Table 1, the correlation between tissues was strong and positive, most were between 0.75 – 0.85. Comparing with previous research in great tit (Parus major), where a positive correlation between 0.6 – 0.8 was found in the methylation level between tissues from blood, liver, gonad and hypothalamus (Lindner et al. 2021), the correlation found in this research was higher, suggesting that the DNA methylation level in the blood of house sparrow should be able to represent the DNA methylation level in other tissues. In addition, it is worth mentioning that the methylation level in blood were lower than other tissues both in this

(20)

19

research and in previous research in European starling (Siller and Rubenstein 2019) and great tit (Derks et al. 2016).

Secondly, there was a difference between sexes in the distribution of DNA methylated sites as suggested in Table 2 and Fig.3. Males had more DNA hypermethylated sites in brain and blood, while females had more in gonads, and the difference in liver was not significant.

Given that differences in DNA methylation has been linked to phenotype variation (Leenen, Muller, and Turner 2016), this sex and tissue specific pattern suggested that many of the phenotypic differences between sexes could be due to DNA methylation variation in gonads, brains and blood rather than in livers. However, as suggested in some recent research (Héberlé and Bardet 2019), this classical view of the DNA methylation machinery may be interpreted in a different way. DNA methylation on the promoter region of a specific gene could be the result of binding between a transcription factor and silencer region (Zhu, Wang, and Qian 2016), which severs as the signal for DNA methylation.

Plus, about 40 % of the differentially methylated sites were found within genes or their mRNA and approximately 20 % of them were found in the coding regions or the exons of these genes or mRNA. Because a gene is composed of exon, intron and regulatory regions (e.g., promoter, enhancer, etc.), this result suggested that around 20 % of the differentially methylated sites were located in the introns or regulatory regions of genes. DNA methylation difference in the regulatory regions like promoter (Siller and Rubenstein 2019) of the gene could potentially lead to expression difference. I indeed found some of the regulatory region

(21)

20

in this study, which is the untranslated region (UTR) on the 3’ or 5’ end of a mRNA. These UTRs were previously known to have an important role in post-transcriptional regulation, regulating the efficiency of translation, stability, etc. (Mignone et al. 2002). But since we currently lack the research on the location of regulatory regions, including UTRs, most of their positions were not provided in the reference genome annotation this research used, details regarding the difference those differently methylated sites can bring was unclear in this study. In addition, some research suggested that the tissue-specific DNA methylation pattern are ubiquitous across species (Wan et al. 2015) and some of those differentially methylated regions could perhaps be conserved in different species in mammals (Zhou et al. 2017), in which the study suggested that 11 % of the differentially methylated regions in human blood and rat blood was conserved, and 27 % of those regions were conserved between different species in rats. Further study could be done across different bird species to see if that’s the cross-species conservation exists in house sparrow and other birds. Besides, in some research, DNA methylation was found to have a functional link with other gene expression regulatory mechanisms, for example, histone lysine methylation (Rose and Klose 2014). DNA

methylation can help target the region needed histone lysine methylation, and vice versa (Rose and Klose 2014). Therefore, not only directly lead to phenotypic variance between sexes, those differentially methylated sites I found in this study can also reflect the pattern of other epigenetic modification in house sparrow.

Thirdly, for the differentially methylated gene between sexes, the result of the annotation comparison was also different between tissues. Most of the sex differentially methylated sites

(22)

21

were found in chromosome Z as expected. In a study on the methylation pattern in chicken, researchers found some region called “male hypermethylated region (MHM)” on

chromosome Z (Teranishi et al. 2001), and some of these male-biased regions were consistent across tissues (Sun et al. 2019). Due to this reason, I will focus on those matched gene

annotation results in Z chromosome here. For genes on the house sparrow Z chromosome, two matches were found out of 970 matches in brain, eight matches were found out of 1114 matches in liver, 18 matches were observed out of 3655 matches in gonad, while no match was found in blood. This difference between tissues, especially the failure to find any chromosome Z gene match in the blood sample, suggested that the methylation pattern was indeed highly different among tissues in house sparrow. Though blood samples could reflect the methylation level in other tissues as suggested above, they might not be ideal material for further analysis in the function of those methylated sites. Using blood sample to test the methylation profile in other tissues may give misleading result as suggested in previous research (Husby 2020).

Though most of the differentially methylated sites between sexes were located to Z chromosome, not too many matches were related to it as mentioned above. For gonads, unexpectedly, the top numbers of matched gene region were found in chromosome 5 (327 matches), chromosome 8 (227 matches) and chromosome 9 (194 matches). Similar ranks were found in other tissues, where few matches were found in Z chromosome despite the high proportion of differentially methylated sites in Z chromosome and chromosome 5 took the leads in the number of matches. The reason why this gap between the number of sexes

(23)

22

differentially methylated sites and the gene annotation matches existed is unknown, but it may be due to the lack of understanding in the genes on chromosome Z. It is likely that many of these sites located in non-coding regions on the genome, in which DNA methylation difference can still lead to expression variation (Yu et al. 2020).

Given the limited time of this research, I only compared the sites of differentially methylated between sexes with the house sparrow gene annotation for their genomic location, no

functional analysis was completed. For further studies, a gene ontology (GO) enrichment could be done on some of the genes matched to see if there is any pathway that can explain the tissue and sex specific difference found above. One of the potential gene for sex-specific difference analysis was a gene located on Z chromosome, differentially methylated site was found within the mRNA region of it. This gene is a similar to chicken (Gallus gallus) gene RMI1, whose product is RecQ-mediated genome instability protein 1. This protein is an essential component of the RMI complex, which has an important role in processing homologous recombination intermediates during the DNA crossover formation in cells (Hoadley et al. 2012). In another words, it is linked to cell proliferation, therefore, the difference in cell types in gonads of males and females could be the reason why this site was differently methylated. This site had a higher percentage of samples that were methylated in female ovary, suggesting that its expression in female was suppressed, which could be explained by the lower need of cell proliferation in ovary due to the longer and lower production of mature ova. In the contrary, the overwhelmingly higher production of sperms might lead to the higher expression of RMI1 in male house sparrow. Further gene ontology

(24)

23

(GO) analysis could be conducted on this gene to determine if that is the case in house sparrow.

(25)

24

5. Conclusion

This work demonstrates sex and tissue specific difference in the DNA methylation pattern in the house sparrow, with higher similarity in homologous tissue compared to in heterologous tissue. A clear difference of DNA methylation pattern between sexes in the same kinds of tissue can only be found in gonads. The methylation levels in all four tissues were positively correlated and ranged from 0.75 to 0.85, suggesting that the methylation level tested using blood sample would be a fine representation to the level in other tissues. Most of the differently methylated sites between sexes were found in gonads and located to the Z chromosome and the fewest sex differences were detected in the blood. A total of 5971 out of 9831 differentially methylated sites between sexes were found to be within genes. Around 40 % of them were within the exon or coding region while most of the other 20 % was within the other parts of a gene. Further study in gene ontology can be done to find the potential pathway leading to sex and tissue specific differences.

(26)

25

Acknowledgement

I would like to give my most sincere thanks to my supervisor Dr. Arild Husby, for all his kind guidance and support on this project, which I am sure will be helpful in my future study and career. I would also like to thank Gabriel David for his support and help in programming and making some of the data files I needed. To Hamish Andrew Burnett in NTNU, I would like to offer my gratitude in sending me his R script for SV annotation match selection. And thanks to all the teachers, assistants, and fellow students I met in Uppsala University, for rekindling my passion and enthusiasm towards scientific research.

(27)

26

Reference

Akalin, A., M. Kormaksson, S. Li, F. E. Garrett-Bakelman, M. E. Figueroa, A. Melnick, and C.

E. Mason. 2012. 'methylKit: a comprehensive R package for the analysis of genome- wide DNA methylation profiles', Genome Biology, 13.

Andrew, S. 2018. "FastQC: a quality control tool for high throughput sequence data.

https://qubeshub.org/resources/fastqc." In.

Bock, C., I. Beerman, W. H. Lien, Z. D. Smith, H. C. Gu, P. Boyle, A. Gnirke, E. Fuchs, D. J.

Rossi, and A. Meissner. 2012. 'DNA Methylation Dynamics during In Vivo Differentiation of Blood and Skin Stem Cells', Molecular Cell, 47: 633-47.

Boyle, P., K. Clement, H. C. Gu, Z. D. Smith, M. Ziller, J. L. Fostel, L. Holmes, J. Meldrim, F.

Kelley, A. Gnirke, and A. Meissner. 2012. 'Gel-free multiplexed reduced representation bisulfite sequencing for large-scale DNA methylation profiling', Genome Biology, 13.

Cubas, Pilar, Coral Vincent, and Enrico Coen. 1999. 'An epigenetic mutation responsible for natural variation in floral symmetry', Nature, 401: 157-61.

Deaton, A. M., and A. Bird. 2011. 'CpG islands and the regulation of transcription', Genes &

Development, 25: 1010-22.

Derks, M. F. L., K. M. Schachtschneider, O. Madsen, E. Schijlen, K. J. F. Verhoeven, and K.

van Oers. 2016. 'Gene and transposable element methylation in great tit (Parus major) brain and blood', Bmc Genomics, 17.

Diedenhofen, Birk, and Jochen Musch. 2015. 'cocor: A Comprehensive Solution for the Statistical Comparison of Correlations', Plos One, 10: e0121945.

Elder, E., and S. McGraw. 2020. 'Comprehensive spatiotemporal DNA methylation analysis of

(28)

27

mouse tissue and organ progression through fetal development to adulthood', Biology of Reproduction, 103: 915-17.

Héberlé, Éléa, and Anaïs Flore Bardet. 2019. 'Sensitivity of transcription factors to DNA methylation', Essays in biochemistry, 63: 727-41.

Herrera, C. M., M. I. Pozo, and P. Bazaga. 2012. 'Jack of all nectars, master of most: DNA methylation and the epigenetic basis of niche width in a flower-living yeast', Molecular Ecology, 21: 2602-16.

Hoadley, Kelly A., Yutong Xue, Chen Ling, Minoru Takata, Weidong Wang, and James L. Keck.

2012. 'Defining the molecular interface that connects the Fanconi anemia protein FANCM to the Bloom syndrome dissolvasome', Proceedings of the National Academy of Sciences, 109: 4437.

Husby, Arild. 2020. 'On the Use of Blood Samples for Measuring DNA Methylation in Ecological Epigenetic Studies', Integrative and Comparative Biology, 60: 1558-66.

Kang, S. W., M. Madkour, and W. J. Kuenzel. 2017. 'Tissue-Specific Expression of DNA Methyltransferases Involved in Early-Life Nutritional Stress of Chicken, Gallus gallus', Frontiers in Genetics, 8.

Kilvitis, H. J., A. W. Schrey, A. K. Ragsdale, A. Berrio, S. M. Phelps, and L. B. Martin. 2019.

'DNA methylation predicts immune gene expression in introduced house sparrows Passer domesticus', Journal of Avian Biology, 50.

Krueger, Felix 2017. 'Trim Galore (RRID:SCR_011847)',

https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/.

Krueger, Felix, and Simon R. Andrews. 2011. 'Bismark: a flexible aligner and methylation

(29)

28

caller for Bisulfite-Seq applications', Bioinformatics, 27: 1571-72.

Kundaje, A., W. Meuleman, J. Ernst, M. Bilenky, A. Yen, A. Heravi-Moussavi, P. Kheradpour, Z. Zhang, J. Wang, M. J. Ziller, V. Amin, J. W. Whitaker, M. D. Schultz, L. D. Ward, A.

Sarkar, G. Quon, R. S. Sandstrom, M. L. Eaton, Y. C. Wu, A. R. Pfenning, X. Wang, M.

Claussnitzer, Y. Liu, C. Coarfa, R. A. Harris, N. Shoresh, C. B. Epstein, E. Gjoneska, D.

Leung, W. Xie, R. D. Hawkins, R. Lister, C. Hong, P. Gascard, A. J. Mungall, R. Moore, E. Chuah, A. Tam, T. K. Canfield, R. S. Hansen, R. Kaul, P. J. Sabo, M. S. Bansal, A.

Carles, J. R. Dixon, K. H. Farh, S. Feizi, R. Karlic, A. R. Kim, A. Kulkarni, D. Li, R.

Lowdon, G. Elliott, T. R. Mercer, S. J. Neph, V. Onuchic, P. Polak, N. Rajagopal, P. Ray, R. C. Sallari, K. T. Siebenthall, N. A. Sinnott-Armstrong, M. Stevens, R. E. Thurman, J. Wu, B. Zhang, X. Zhou, A. E. Beaudet, L. A. Boyer, P. L. De Jager, P. J. Farnham, S.

J. Fisher, D. Haussler, S. J. M. Jones, W. Li, M. A. Marra, M. T. McManus, S. Sunyaev, J. A. Thomson, T. D. Tlsty, L. H. Tsai, W. Wang, R. A. Waterland, M. Q. Zhang, L. H.

Chadwick, B. E. Bernstein, J. F. Costello, J. R. Ecker, M. Hirst, A. Meissner, A.

Milosavljevic, B. Ren, J. A. Stamatoyannopoulos, T. Wang, M. Kellis, and Consortium Roadmap Epigenomics. 2015. 'Integrative analysis of 111 reference human epigenomes', Nature, 518: 317-30.

Laine, Veronika N., Toni I. Gossmann, Kyle M. Schachtschneider, Colin J. Garroway, Ole Madsen, Koen J. F. Verhoeven, Victor de Jager, Hendrik-Jan Megens, Wesley C. Warren, Patrick Minx, Richard P. M. A. Crooijmans, Pádraic Corcoran, Frank Adriaensen, Eduardo Belda, Andrey Bushuev, Mariusz Cichon, Anne Charmantier, Niels Dingemanse, Blandine Doligez, Tapio Eeva, Kjell Einar Erikstad, Slava Fedorov,

(30)

29

Michaela Hau, Sabine Hille, Camilla Hinde, Bart Kempenaers, Anvar Kerimov, Milos Krist, Raivo Mand, Erik Matthysen, Reudi Nager, Claudia Norte, Markku Orell, Heinz Richner, Tore Slagsvold, Vallo Tilgar, Joost Tinbergen, Janos Torok, Barbara Tschirren, Tera Yuta, Ben C. Sheldon, Jon Slate, Kai Zeng, Kees van Oers, Marcel E. Visser, Martien A. M. Groenen, and Consortium The Great Tit HapMap. 2016. 'Evolutionary signals of selection on cognition from the great tit genome and methylome', Nature communications, 7: 10474.

Leenen, F. A. D., C. P. Muller, and J. D. Turner. 2016. 'DNA methylation: conducting the orchestra from exposure to phenotype?', Clinical Epigenetics, 8.

Liebl, A. L., A. W. Schrey, C. L. Richards, and L. B. Martin. 2013. 'Patterns of DNA Methylation Throughout a Range Expansion of an Introduced Songbird', Integrative and Comparative Biology, 53: 351-58.

Lindner, M., I. Verhagen, H. M. Viitaniemi, V. N. Laine, M. E. Visser, A. Husby, and K. van Oers. 2021. 'Temporal changes in DNA methylation and RNA expression in a small song bird: within- and between-tissue comparisons', Bmc Genomics, 22.

Liu, S., K. P. Sun, T. L. Jiang, J. P. Ho, B. Liu, and J. Feng. 2012. 'Natural epigenetic variation in the female great roundleaf bat (Hipposideros armiger) populations', Molecular Genetics and Genomics, 287: 643-50.

Liu, Y. B., P. Siejka-Zielinska, G. Velikova, Y. Bi, F. Yuan, M. Tomkova, C. S. Bai, L. Chen, B.

Schuster-Bockler, and C. X. Song. 2019. 'Bisulfite-free direct detection of 5- methylcytosine and 5-hydroxymethylcytosine at base resolution', Nature Biotechnology, 37: 424-+.

(31)

30

Massicotte, R., E. Whitelaw, and B. Angers. 2011. 'DNA methylation A source of random variation in natural populations', Epigenetics, 6: 422-28.

Mignone, Flavio, Carmela Gissi, Sabino Liuni, and Graziano Pesole. 2002. 'Untranslated regions of mRNAs', Genome Biology, 3: reviews0004.1.

OriginLab Corporation. 2017. 'Origin (Pro), Version 2017', Northampton, MA, USA.

Quinlan, Aaron R., and Ira M. Hall. 2010. 'BEDTools: a flexible suite of utilities for comparing genomic features', Bioinformatics, 26: 841-42.

R Core Team. 2020. R: A language and environment for statistical computing (R Foundation for Statistical Computing, Vienna, Austria).

Radhakrishnan, S., R. Literman, J. L. Neuwald, and N. Valenzuela. 2018. 'Thermal Response of Epigenetic Genes Informs Turtle Sex Determination with and without Sex Chromosomes', Sexual Development, 12: 308-19.

Rose, N. R., and R. J. Klose. 2014. 'Understanding the relationship between DNA methylation and histone lysine methylation', Biochim Biophys Acta, 1839: 1362-72.

Satoh, M. 1991. 'Histogenesis and organogenesis of the gonad in human embryos', Journal of anatomy, 177: 85-107.

Schiefelbein, S. H. H., A. Kamal, Z. She, A. Rentmeister, and H. B. Kraatz. 2018. 'Direct Bisulfite-Free Detection of 5-Methylcytosine by Using Electrochemical Measurements Aided by a Monoclonal Antibody', Chemelectrochem, 5: 1631-35.

Schlichting, C. D., and M. A. Wund. 2014. 'PHENOTYPIC PLASTICITY AND EPIGENETIC

MARKING: AN ASSESSMENT OF EVIDENCE FOR GENETIC

ACCOMMODATION', Evolution, 68: 656-72.

(32)

31

Schrey, A. W., M. Alvarez, C. M. Foust, H. J. Kilvitis, J. D. Lee, A. L. Liebl, L. B. Martin, C.

L. Richards, and M. Robertson. 2013. 'Ecological Epigenetics: Beyond MS-AFLP', Integrative and Comparative Biology, 53: 340-50.

Siller, S. J., and D. R. Rubenstein. 2019. 'A Tissue Comparison of DNA Methylation of the Glucocorticoid Receptor Gene (Nr3c1) in European Starlings', Integrative and Comparative Biology, 59: 264-72.

Sun, Dan, Donna L. Maney, Thomas S. Layman, Paramita Chatterjee, and Soojin V. Yi. 2019.

'Regional epigenetic differentiation of the Z Chromosome between sexes in a female heterogametic system', Genome research, 29: 1673-84.

Suzuki, M. M., and A. Bird. 2008. 'DNA methylation landscapes: provocative insights from epigenomics', Nature Reviews Genetics, 9: 465-76.

Teranishi, M., Y. Shimada, T. Hori, O. Nakabayashi, T. Kikuchi, T. Macleod, R. Pym, B.

Sheldon, I. Solovei, H. Macgregor, and S. Mizuno. 2001. 'Transcripts of the MHM region on the chicken Z chromosome accumulate as non-coding RNA in the nucleus of female cells adjacent to the DMRT1 locus', Chromosome Res, 9: 147-65.

Wan, Jun, Verity F. Oliver, Guohua Wang, Heng Zhu, Donald J. Zack, Shannath L. Merbs, and Jiang Qian. 2015. 'Characterization of tissue-specific differential DNA methylation suggests distinct modes of positive and negative gene expression regulation', Bmc Genomics, 16: 49.

Wang, H. Q., L. K. Tuominen, and C. J. Tsai. 2011. 'SLIM: a sliding linear model for estimating the proportion of true null hypotheses in datasets with dependence structures', Bioinformatics, 27: 225-31.

(33)

32

Xiao, Xinlong, Jieqiong Zhang, Tao Li, Xing Fu, Viswanathan Satheesh, Qingfeng Niu, Zhaobo Lang, Jian-Kang Zhu, and Mingguang Lei. 2019. 'A group of SUVH methyl-DNA binding proteins regulate expression of the DNA demethylase ROS1 in Arabidopsis', Journal of Integrative Plant Biology, 61: 110-19.

Yong, W. S., F. M. Hsu, and P. Y. Chen. 2016. 'Profiling genome-wide DNA methylation', Epigenetics & Chromatin, 9.

Yu, Liyuan, Kai Xia, Xiao Cen, Xinqi Huang, Wentian Sun, Zhihe Zhao, and Jun Liu. 2020.

'DNA methylation of noncoding RNAs: new insights into osteogenesis and common bone diseases', Stem Cell Research & Therapy, 11: 109.

Zaret, K. S. 2008. 'Genetic programming of liver and pancreas progenitors: lessons for stem- cell differentiation', Nature reviews. Genetics, 9: 329-40.

Zhou, Jia, Renee L. Sears, Xiaoyun Xing, Bo Zhang, Daofeng Li, Nicole B. Rockweiler, Hyo Sik Jang, Mayank N. K. Choudhary, Hyung Joo Lee, Rebecca F. Lowdon, Jason Arand, Brianne Tabers, C. Charles Gu, Theodore J. Cicero, and Ting Wang. 2017. 'Tissue- specific DNA methylation is conserved across human, mouse, and rat, and driven by primary sequence conservation', Bmc Genomics, 18: 724.

Zhu, Heng, Guohua Wang, and Jiang Qian. 2016. 'Transcription factors as readers and effectors of DNA methylation', Nature reviews. Genetics, 17: 551-65.

(34)

33

Appendix

1. R script for clustering, PCA and finding differently methylated region

library(methylKit)

setwd("D:/DegreeProject/Bismarkfiles/") #set working directory, no blank in the path

all.id <- list("8M71323-B","8N06837-B","8N72676-B","8N72694-B",…) #sample id all.treatment <- list(1,1,0,1,…) #sample treatment, 1 for males, 0 for females as control

all.path <- list("8M71323-B_S44_L008_R1_001.bismark.cov","8N06837- B_S1_L001_R1_001.bismark.cov","8N72676-

B_S47_L008_R1_001.bismark.cov","8N72694-B_S41_L007_R1_001.bismark.cov",…)

#importing file path

AT <- methRead(all.path, sample.id = all.id, assembly="none", dbtype =

"tabix", treatment=all.treatment, header=FALSE, pipeline="bismarkCoverage")

#read all the file, store the result in tabix to avoid memory crush

FAT <- filterByCoverage(AT,lo.count=10,lo.perc=NULL,

hi.count=NULL,hi.perc=99.9) #filter AT with 10x coverage, discard regions with coverage higher than

99.9 %

UAT <- unite(FAT, destrand=FALSE) #unite the filtered files into one for further analysis

UAT.filtered <- select(UAT, 1:81531) #select rows covering sparrow chromosome (chr name start with CM), UAT has 93498 rows, 1 to 81531 is CM, 81532 to 93498 is MBAE (unplaced clips)

clusterSamples(UAT.filtered,dist="correlation", method="ward", plot=TRUE)

#print cluster figure (Fig.1), cluster based on the correlation factor between samples (or any of the following,

(35)

34

"euclidean", "maximum", "manhattan", "canberra", "binary" or "minkowski"), using method ward (or any of the following, "ward.D", "ward.D", "single", "complete", "average", "mcquitty", "median" or "centroid")

cor.mat = cor(percMethylation(UAT.filtered)) #produce a matrix of correlation coefficients for analysis

PCA <- PCASamples(UAT.filtered, screeplot = FALSE) #print PCA plot (Fig.2), calculation

result is stored in “PCA”, can use the “standard deviations” provided in “PCA” to calculate

AT.filtered.diff <- calculateDiffMeth(UAT.filtered) #calculate methylation difference AT.diff.filtered <- getMethylDiff(AT.filtered.diff,difference=15,qvalue=0.01,type =

all) #get the calculation result, difference should be higher than 15 %, q-value should be lower than 0.01, both

hyper- and hypomethylated results are returned (or “hyper”, “hypo”)

write.table(AT.diff.filtered,"D:/DegreeProject/Bismarkfiles/ATDF.csv",row.n

ames=FALSE,col.names=TRUE,sep=",") #export the result to csv file for annotation analysis

AT.f.diff.per.15 <-

diffMethPerChr(AT.filtered.diff,plot=TRUE,qvalue.cutoff=0.01,

meth.cutoff=15) #print the methylation difference percentage result for q-value < 0.01, methylation

difference > 15%, plot = TRUE return Fig.3, plot =FALSE return Table 1

#For analysis in different tissue, need to import the specific tissue files and run the whole process again

2. BEDtools command in finding matches in annotation reference genome

bedtools intersect -u -wa -a (genome annotation file path)/GCA_001700915.1_Passer_domesticus-1.0_genomic_annotation.gff -

b (differently methylated sites file).bed

Sex and tissue specific DNA methylationpatterns in the house sparrow (Passerdomesticus)Yuming Shi