Comparative analysis reveals distinctive epigenetic features of the human cerebellum.

Identifying the molecular underpinnings of the neural specializations that underlie human cognitive and behavioral traits has long been of considerable interest. Much research on human-specific changes in gene expression and epigenetic marks has focused on the prefrontal cortex, a brain structure distinguished by its role in executive functions. The cerebellum shows expansion in great apes and is gaining increasing attention for its role in motor skills and cognitive processing, including language. However, relatively few molecular studies of the cerebellum in a comparative evolutionary context have been conducted. Here, we identify human-specific methylation in the lateral cerebellum relative to the dorsolateral prefrontal cortex, in a comparative study with chimpanzees (Pan troglodytes) and rhesus macaques (Macaca mulatta). Specifically, we profiled genome-wide methylation levels in the three species for each of the two brain structures and identified human-specific differentially methylated genomic regions unique to each structure. We further identified which differentially methylated regions (DMRs) overlap likely regulatory elements and determined whether associated genes show corresponding species differences in gene expression. We found greater human-specific methylation in the cerebellum than the dorsolateral prefrontal cortex, with differentially methylated regions overlapping genes involved in several conditions or processes relevant to human neurobiology, including synaptic plasticity, lipid metabolism, neuroinflammation and neurodegeneration, and neurodevelopment, including developmental disorders. Moreover, our results show some overlap with those of previous studies focused on the neocortex, indicating that such results may be common to multiple brain structures. These findings further our understanding of the cerebellum in human brain evolution.

Introduction

Humans’ remarkable cognitive abilities–enabling language use, complex technology, and cultural behavior—are hallmarks of our species. The neurobiological underpinning of these traits originate in large part from developmentally established species-specific spatial and temporal patterns of gene regulation in the brain [1-6]. One important gene regulatory mechanism critical to establishing species-typical and cell type-specific transcriptional profiles is CpG methylation. This epigenetic mark involves the addition of a methyl chemical group by methyltransferase enzymes to cytosine DNA bases occurring next to guanine bases, or CpG sites, and can affect gene expression by influencing transcription factor binding and chromatin organization [7]. Methylation across the genome can be considered to represent the footprint of developmentally configured, cell type-specific gene regulatory settings [8]. Consistent species differences in methylation may thus reflect evolved developmental differences.

As such, methylation, along with gene expression, has been studied in a comparative evolutionary context in an effort to identify clues to the molecular basis of human-specific traits. Indeed, differential methylation among closely related primate species has been identified in several tissues, including kidney, liver, blood, and bone [9-12]. Comparing brain methylation patterns among humans and other primates to reveal species differences in developmental programming and plasticity is thus of considerable interest in the search for what sets human cognition apart. Although nonhuman primates share the same brain structures as humans, some structures have taken on new roles that enable human-specific cognitive abilities and behaviors, which may be reflected in changes in relative size, microstructure, or connectivity [13]. For example, the question of whether humans have a relatively enlarged prefrontal cortex, a region that plays a role in language and abstract thought, has been the subject of much study, and the connectivity and neurotransmitter innervation of this structure appears to be distinct in humans [13,14].

Such changes in our lineage presumably have molecular bases and previous studies have found differential methylation between humans and chimpanzees in the prefrontal cortex associated with genes involved in developmental processes, psychiatric conditions, and neurodegeneration [15,16], as well as genes implicated in brain size [17] and regulation of the language-associated gene FOXP2 [18]. Together, these findings affirm the potential relevance of methylation in brain tissue for understanding human neurobiology and evolution.

We build on this evidence by comparing genome-wide methylation in humans, chimpanzees (Pan troglodytes), and rhesus macaques (Macaca mulatta) in the dorsolateral prefrontal cortex (DLPFC), as well as the lateral cerebellum (Fig 1). The DLPFC is a specialized part of the prefrontal cortex involved in executive functions, which are important for working memory, attention, cognitive flexibility, and planning (for review, see [19]). It is also heavily interconnected with language areas of the cortex and striatum in humans [20]. Moreover, the human DLPFC is distinguished from other primates by patterns of cell type composition and distribution, as well as neurochemistry [21,22]. The DLPFC is also especially affected in autism spectrum disorder and schizophrenia [19]. The DLPFC has long been a focus of comparative evolutionary studies, including molecular analyses.

Fig 1

Species relationships, brain structures, and sample clustering.

A = Phylogenetic tree of species included in this study from TimeTree.org [113], B = brain structures included in this study depicted on a chimpanzee brain illustration modified from [114], C and D = hierarchical clustering of samples from the dorsolateral prefrontal cortex (C) and cerebellum (D) based on genome-wide methylation after filtering and normalizing using correlations between samples as distance. DLPFC = dorsolateral prefrontal cortex.

In contrast, the cerebellum has received comparatively little attention in human evolutionary neuroscience until fairly recently. Long recognized for its role in motor control, later discoveries of its role in non-motor function, including cognitive processes like language [23,24], spurred a surge of research on the cerebellum [25,26]. Intriguingly, comparative work has revealed that the lateral cerebellum is particularly enlarged in humans and other great apes [27-29], and its claim to the majority of the brain’s neurons [30] has furthered interest in the cerebellum’s role in human evolution. Indeed, it is increasingly recognized that the cerebellum likely plays a role in the fine motor skills necessary for speech and the precision and control underlying many cultural behaviors distinguishing humans [26]. Additionally, it is involved in cognitive processing of language [31,32] and executive function, with the cerebellum potentially allowing for a kind of supervised machine learning on spatiotemporal data input to build cognitive models and automate complex behaviors [25].

Nevertheless, comparative molecular studies have continued to focus predominately on the neocortex. In the current study, we thus sought to characterize epigenetic differences among humans and other species in the cerebellum to gain insights into potential human distinctiveness, as well as provide context to comparative studies of the neocortex. We specifically investigated the lateral cerebellum, given its enlargement in hominoids and extensive interconnection with the association areas of the neocortex [29].

Here, we examined species differences in methylation in the DLPFC and lateral cerebellum using a fairly large sample size of rare nonhuman primate samples and a large number (> 160,000) of CpG sites across the genome. Specifically, we identified human-specific differentially methylated genomic regions specific to each brain structure, employing a three-way comparison to identify whether differences between humans and chimpanzees occurred along the chimpanzee or human branch. We also determined whether species-specific patterns of methylation are unique to each brain structure or show overlap among structures or with blood, a non-neural tissue. We were especially interested in differentially methylated regions (DMRs) associated with genes and covering likely regulatory elements (e.g., near gene transcription start sites). We further determined whether genes associated with such regulatory-associated DMRs (regDMRs) showed species differences in gene expression.

Results

Differential methylation analysis

We analyzed DLPFC and cerebellum samples from seven human, eight chimpanzee, and seven rhesus macaque adults free of neuropathology (Methods; Fig 1A and S1 Table). We extracted DNA from 0.03–0.1 grams of tissue using the Qiagen DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany), brought DNAs to a standard concentration (~70 ng/μl), and profiled genome-wide methylation levels from bisulfite-converted DNA by microarray (Illumina Infinium EPIC Methylation BeadChip; “EPIC array” hereafter). We retained only CpG sites expected to be efficiently captured across all three species in downstream analysis and used surrogate variable analysis to account for cell type heterogeneity (Methods). We identified human and chimpanzee-specific differentially methylated regions (DMRs) comprised of at least five CpGs, including at least three that showed an absolute difference in the proportion of methylation of at least 0.15 from the other two species. We identified several significant human- and chimpanzee-specific DMRs in each of the two brain structures (Table 1). Both species showed a greater number of DMRs in the cerebellum than in the DLPFC (human: DLPFC = 28, cerebellum = 64; chimpanzee: DLPFC = 5, cerebellum: 40). Many human- and brain structure-specific DMRs overlapped genes involved in processes and conditions of potential relevance to aspects of human-specific neurobiology, such as neurodevelopment, including neurodevelopmental disorders (Figs 2 and 3).

Table 1

Human and chimpanzee brain region-specific regDMRs.

Brain structure	Species	Gene	Gene name	Hyper or hypo*
DLPFC	Human	PAQR4 KREMEN2	progestin and adipoQ receptor family member 4 kringle containing transmembrane protein 2	hyper
		MYLK	myosin light chain kinase	hypo
		ADAM30	ADAM metallopeptidase domain 3	mix
		MLST8	MTOR associated protein, LST8 homolog	hypo
		TMEM177	transmembrane protein 177	hyper
		TRIM65	tripartite motif containing 65	hyper
		C10orf71 & C10orf71-AS1	chromosome 10 open reading frame 71	hyper
		FAM193A	family with sequence similarity 193 member A	hypo
		SLC13A4	solute carrier family 13 member 4	hyper
	Chimp-anzee	LRCH1	leucine rich repeats and calponin homology domain containing 1	hyper
		CUGBP1	CUGBP Elav-like family member 1	hyper
Cerebellum	Human	GP5	glycoprotein V platelet	hyper
		CHAD ACSF2	chondroadherin acyl-CoA synthetase family member 2	hypo
		SEPP1	selenoprotein P	hyper
		DLEU7	deleted in lymphocytic leukemia 7	hyper
		HRH1	histamine receptor H1	hyper
		PARVG	parvin gamma	hypo
		TRAF3IP2	TRAF3 interacting protein 2	hypo
		SH3BGR	SH3 domain binding glutamate rich protein	hyper
		FAM83A	family with sequence similarity 83 member A	hypo
		ANKS1B	ankyrin repeat and sterile alpha motif domain containing 1B	hypo
		DLGAP1 & DLGAP1-AS1/FLJ35776	DLG associated protein 1	hypo
		PLCH1	phospholipase C eta 1	hypo
		FAM198B	golgi associated kinase 1B	hypo
		TRAF3IP2	TRAF3 interacting protein 2	hypo
		STK33	serine/threonine kinase 33	hypo
		CCDC8	coiled-coil domain containing 8	hypo
		GSTO2	glutathione S-transferase omega 2	hypo
		CRIP3	cysteine rich protein 3	mix
		CPNE6	copine 6	hypo
		PHACTR1	phosphatase and actin regulator 1	hypo
		FUT4	fucosyltransferase 4	hypo
		MEDAG	mesenteric estrogen dependent adipogenesis	mix
		MIR192 & MIR194-2	microRNA 192microRNA 194–2	hypo
		MIR1182 & FAM89A	microRNA 1182 family with sequence similarity 89 member A	hyper
		DNAH10	dynein axonemal heavy chain 10	hyper
		SDR42E1	short chain dehydrogenase/reductase family 42E, member 1	hyper
		LOC101928322	CELF2 divergent transcript	hypo
		DMPK	DM1 protein kinase	hyper
	Chimp-anzee	CCDC140	CCDC140 long non-coding RNA	hyper
		CCDC102A	coiled-coil domain containing 102A	hyper
		LINGO3	leucine rich repeat and Ig domain containing 3	hypo
		ZNF608	Zinc finger protein 68	hyper
		MYO18A	myosin XVIIIA	hypo
		CDK15	cyclin dependent kinase 15	hypo
		TRIM66	tripartite motif containing 66	hyper
		SLC43A3	solute carrier family 43 member 3	hypo
		ADAM32	ADAM metallopeptidase domain 32	hypo
		SYTL1	synaptotagmin like 1	hyper
		TCF21	transcription factor 21	hyper
		BNC1	basonuclin 1	hyper
		RNPEPL1	arginyl aminopeptidase like 1	hypo
		APPL2	adaptor protein, phosphotyrosine interacting with PH domain and leucine zipper 2	hyper
		GYPC	glycophorin C (Gerbich blood group)	hyper
		DRD2	dopamine receptor D2	hypo
		NODAL	nodal growth differentiation factor	hyper
		ST3GAL1	ST3 beta-galactoside alpha-2,3-sialyltransferase 1	hyper
		LRRC8D	leucine rich repeat containing 8 VRAC subunit D	hyper
		TWIST2	twist family bHLH transcription factor 2	hyper
		AVEN & CHRM5	apoptosis and caspase activation inhibitor cholinergic receptor muscarinic 5	hypo
		LPL	lipoprotein lipase	hypo
		KCNE3	potassium voltage-gated channel subfamily E regulatory subunit 3	hyper

*Hypermethylated or hypomethylated

Fig 2

DLPFC-specific human regDMRs associated with ADAM30 and FAM193A.

A = ADAM30, B = FAM193A. Lines are loess smoothed methylation values for each species across the DMR and each point represents raw methylation values for each individual at each CpG site within the DMR ranges. TSS200 = within 200 bp of a transcription start site and TSS1500 within 1500 bp of a transcription start site. Chromosomal coordinates are in base pairs and refer to human genome build hg19.

Fig 3

Cerebellum-specific human regDMRs associated with PLCH1, ANKS1B, DLGAP1, and PHACTR1.

A = PLCH1, B = ANKS1B, C = DLGAP1, D = PHACTR1. Lines are loess smoothed methylation values for each species across the DMR and each point represents raw methylation values for each individual at each CpG site within the DMR ranges. TSS200 = within 200 bp of a transcription start site and TSS1500 within 1500 bp of a transcription start site. Chromosomal coordinates are in base pairs and refer to human genome build hg19.

Brain structure specificity and regulatory element association

Some (~7%) of the DMRs identified were evident in both brain structures (e.g., S1 Fig). We also compared our DMR dataset with a dataset of chimpanzee-human DMRs in blood. Of the DLPFC DMRs, ~27% were also identified in blood, and, of the cerebellum DMRs, ~7% were also called as DMRs in blood. Five DMRs were shared across both brain structures and blood (e.g., S2 Fig). Overall, correspondence across tissues was high, with average global DLPFC and cerebellum methylation showing a correlation of 0.91–0.92 in all three species, DLPFC and blood methylation showing a correlation of 0.90–0.91 in humans and chimpanzees, and cerebellum and blood showing a correlation of ~0.82 in humans and chimpanzees, the two species represented in the blood dataset.

We next determined whether DMRs were associated with potential regulatory regions (Methods). In the DLPFC, nine human-specific and two chimpanzee-specific DMRs unique to the DLPFC were regulatory-associated DMRs (regDMRs; i.e., covering potential regulatory elements; see Methods; S2 Table and Figs 2 and 3). In the cerebellum, we identified 28 such regDMRs that were human-specific, which is over three times as many as in the DLPFC, and 23 that were chimpanzee-specific (Tables 1 and S2).

Correspondence with gene expression and previous studies

We also identified whether genes associated with species-specific regDMRs showed species-specific differential gene expression using a previously published RNA-seq dataset including humans, chimpanzees, and rhesus macaques (Methods). We identified 281 genes that showed human-specific expression in the DLPFC and 487 that showed human-specific expression in the cerebellum. We also identified 231 genes that showed chimpanzee-specific expression in the DLPFC and 339 that showed chimpanzee-specific expression in the cerebellum. In the DLPFC, seven genes associated with human-specific, DLPFC-specific regDMRs were represented in the gene expression dataset. None showed differential gene expression in humans. Two genes associated with chimpanzee-specific, DLPFC-specific regDMRs were represented in the gene expression dataset and neither was associated with differential gene expression. In contrast, in the cerebellum, five (CHAD, STK33, CCDC8, CPNE6, SDR42E1) of the 24 genes associated with human-specific, cerebellum-specific regDMRs that were represented in the gene expression dataset showed significant human-specific differential expression (Fig 4 and S2 Table), which represents more than would be expected by chance (Fisher’s exact test; p = 0.004). Four of these genes show the canonical inverse relationship between methylation and gene expression, with three (CHAD, STK33, CPNE6) showing hypomethylation and upregulation (Fig 4) and one (SDR42E1) showing hypermethylation and downregulation. The fifth gene (CCDC8) unexpectedly shows the opposite pattern with hypermethylation and upregulation. Eighteen genes associated with chimpanzee-specific, cerebellum-specific regDMRs were represented in the gene expression data and two (SYTL1 and NODAL) showed significant, chimpanzee-specific differential gene expression (S3 Fig and S2 Table). This degree of overlap is not more than what would be expected by chance (Fisher’s exact test; p = 0.12). Both show inverse methylation and gene expression with SYTL1 exhibiting hypomethylation and upregulation and NODAL hypermethylation and downregulation. Macaques, however, show no expression of NODAL, despite showing lower methylation than chimpanzees.

Fig 4

Cerebellum-specific human regDMRs and gene expression levels for CHAD, CPNE6, and STK33.

A = CHAD methylation, B = CHAD gene expression, C = CPNE6 methylation, D = CPNE6 gene expression, E = STK33 methylation, and F = STK33 gene expression. Lines are loess smoothed methylation values for each species across the DMR and each point represents raw methylation values for each individual at each CpG site within the DMR ranges. TSS200 = within 200 bp of a transcription start site and TSS1500 within 1500 bp of a transcription start site. Chromosomal coordinates are in base pairs and refer to human genome build hg19.

We also assessed whether genes showing uniquely human gene expression in the cerebellum shared the gene ontology (GO) term “chemical synaptic transmission” (GO:0007268) that annotated some of our cerebellar regDMR genes (S3 and S4 Tables). We found that 19 genes showing uniquely human gene expression are annotated with this GO term (S3 Table). However, this does not represent statistical enrichment (Fisher’s exact test, p = 0.2763).

Finally, we assessed whether global methylation levels in our dataset showed correspondence with a previously published WGBS dataset for chimpanzees and humans. We found that average methylation levels across sites were highly correlated between our dataset and the WGBS dataset at overlapping CpGs (chimpanzees: Pearson’s R = 0.91; humans: Pearson’s R = 0.90).

Discussion

Our results showing divergence in methylation of the cerebellum comparable to that in the DLPFC is notable giving a long-standing focus on the prefrontal cortex in comparative neuroscience. Moreover, we identified a greater number of human-specific regDMRs in the cerebellum than the DLPFC, and some of these cerebellar human-specific regDMRs were associated with genes that also showed uniquely human gene expression in the cerebellum, which was not the case for the DLPFC. Our finding affirms the importance of the study of the cerebellum in human brain evolution. This finding is consistent with recent analyses of gene expression across many brain structures [33] and the observed disproportionate expansion of the cerebellum in humans [27], as well as in line with increasing recognition of the cerebellum’s potential importance in human-specific cognitive and behavioral traits, including language, speech production, and spatiotemporal processing [25-28,31,32].

Of the five human-specific differentially methylated regions associated with corresponding human-specific gene expression, several are not well characterized or do not have known roles in the brain. CPNE6 (copine 6; Fig 4C and 4D), however, plays a role in synaptic plasticity and learning and, notably, was previously found to show prolonged developmental expression in the human cortex relative to nonhuman primates [34,35].

Several other human-specific regDMRs overlapped genes involved in neurodevelopment (FAM198B, PHACTR1, and DIP2C in the cerebellum) and neurodevelopmental disorders, including autism spectrum disorder (ANKS1B, DIP2C, and DLGAP1 in the cerebellum) and schizophrenia (FAM193A in the DLPFC and ANKS1B, DLGAP1, and DLEU7 in the cerebellum; Figs 2B and 3B and 3C) [36-43]. These results are consistent with previous studies [15,16] and may reflect species differences in development, as well as potentially greater human vulnerability to neuropsychiatric disorders [44-46].

Several genes overlapped by human-specific regDMRs are also implicated in neuroinflammation and neurodegeneration, including ADAM30, GSTO2, PRSS50, MYOZ3, and ANKS1B (Figs 2A and 3B) [47-51]. For example, one of the DLPFC regDMRs spans regulatory elements associated with ADAM30 (ADAM metallopeptidase domain 30; Fig 2A), a gene that plays a role in processing amyloid precursor protein and appears to be downregulated in Alzheimer’s disease [50]. A cerebellum-specific regDMR overlaps regulatory elements associated with ANKS1B (ankyrin repeat and sterile alpha motif domain containing 1B; Fig 3B), which is enriched at synapses in the cerebellum and binds amyloid precursor protein, decreasing amyloid beta protein production [38,48]. Humans may be specifically susceptible to dementia and neurodegeneration with age. Nonhuman primates show cognitive declines [52-55], age-related brain structural changes [56,57], and some degenerative pathology [58-60]. However, these phenotypes are generally considered to be mild compared to the severe degenerative pathology observed in humans [59,61,62]. Greater neurodegeneration in humans is thought to reflect a possible increase in oxidative stress and inflammation resulting from heightened energy metabolism over an extended lifespan [59,63,64].

Notably, a number of the genes associated with regDMRs (CPNE6, DLGAP1, EGFLAM, and ANKS1B) and implicated in neuropathology play roles in synaptic plasticity with several functioning as postsynaptic scaffold components [34,36,48,65]. Dysfunction in synaptic transmission and plasticity during development and adulthood may be a common risk in neurodevelopmental and degenerative disorders [48]. Greater and prolonged synaptic plasticity in humans has been hypothesized to underlie human behavioral and cognitive flexibility [64,66,67]. The cost of this increased, sophisticated cognitive dexterity may thus be another contributor to increased susceptibility to neurodevelopmental and degenerative conditions [66].

Lipid metabolism is also important in the brain as lipids play critical roles in the structure and function of nervous tissue. Notably, we identified a human-specific hypomethylated regDMR in the cerebellum overlapping PLCH1 (Fig 3A), which encodes a phospholipase. PLCH1 was previously found to show human-specific expression and to interact with lipids comprising part of the human-specific brain lipidome profile [68], although we did not identify a human-specific pattern of expression of this gene in the present analysis. Human-specific regDMRs were also associated with several noncoding RNA genes, including LOC101928322, MIR192, and MIR1182. The important roles of noncoding RNAs in the brain are increasingly gaining attention [69].

Chimpanzee-specific DMRs (Tables 1 and S2) were also of interest. Chimpanzees exhibit a cerebellum-specific DMR at DRD2 (S4 Fig), the gene encoding the dopamine receptor D2, one of five receptors for the neurotransmitter dopamine [70]. The genetic modulation of dopaminergic pathways in intra- and interspecific variation in primate social behavior has garnered considerable interest [71-73]; however, the significance of such variation in methylation in the cerebellum requires further study.

Our data contained overall fewer CpGs than a comparative study of the prefrontal cortex in primates using whole genome bisulfite sequencing (WGBS) [15] such that none of the genomic ranges of DMRs identified in that study were represented in our data by enough CpGs (5) to comprise a DMR. We nevertheless found the genomic ranges of the previous study’s human-specific DMRs to be enriched for individual, human-specific differentially methylated CpGs in our datasets for both brain structures (Fisher’s exact test, p < 0.001), with slightly over half the CpGs present in each dataset falling within these ranges showing human-specific methylation. This finding, along with generally strong correlation of methylation levels between datasets across all overlapping CpGs, suggests that our study is broadly concordant with this previous study using a sequencing-based approach. Notably, however, our cerebellum dataset was as enriched for human-specific differentially methylated CpGs in these ranges as was the DLPFC dataset and two genes that we found to show human-specific methylation in the cerebellum (DIP2C and RBPMS) were previously associated with human-specific methylation in the prefrontal cortex [15]. These findings indicate that previously identified human-specific DMRs might not be unique to the prefrontal cortex, or even the neocortex.

Thus, our results suggest that human-specific DMRs previously identified in a particular brain structure may not actually be structure-specific and highlights the value of jointly comparing species and brain structure differences. This is consistent with our findings of substantial (~27%) overlap in human-specific differential methylation between the two brain structures, as well as less extensive, but still notable (~7%) overlap with differential methylation we identified in blood, suggesting that not all differential methylation in the brain necessarily reflects brain-specific regulatory differences. The correlations in global methylation between blood and the DLPFC and cerebellum are highly similar to those previously reported [74], as is the greater correspondence in methylation between blood and neocortex than blood and cerebellum [74,75] in global methylation, as well as DMRs. Overlapping differential methylation with blood might reflect human-specific patterns established early in development in cell types with a common developmental origin. Because there is blood present in the brain, some signal from blood methylation may also be detected in our brain data. Humans, furthermore, may show greater overlap among structures/tissues than chimpanzees because the array targets human functional regions.

Overlapping differential methylation across brain structures might reflect neuron-specific, species-specific patterns of methylation. Along these lines, some degree of divergence in differential methylation that we observed between brain structures might reflect differences in relative cell type composition. Although we have accounted for interindividual differences in cell type composition within regions using a surrogate variable approach, we have not accounted for cell type composition differences between regions. Notably, the cerebellum has a much higher proportion of neurons than the cerebral cortex and the composition of neurons relative to glia varies less among species with brain size [30]. Moreover, compared to the cerebral cortex, there is a higher relative proportion of inhibitory interneurons in the neocortex than the cerebellum [76]. As such, the cerebellum is generally more homogenous in cell composition within and across species, and observed species differences in the cerebellum may reflect neuron- or excitatory neuron-specific rather than region-specific differences. Single cell profiling would address this question.

The modest correspondence with gene expression in our analysis is consistent with the results of previous analyses. Prior studies of species differences in methylation have uncovered that methylation only explains a limited degree of variation in gene expression (e.g., 12–18%), even when gene expression and methylation levels are measured in the same tissue samples from the same individuals [11,15,77], which was not the case in our study. Rather, we leveraged a publicly available gene expression dataset matched for species and brain structure. It is not entirely surprising that gene expression at a given moment only tracks methylation to a limited degree given that methylation is only one of many regulatory mechanisms that influence gene expression and that variation in methylation and gene expression operate on different temporal scales, with methylation being relatively stable and gene expression being characterized by intermittent bursts of transcription, resulting in generally greater fluctuations in gene expression levels than methylation levels [11,78,79]. Furthermore, predicting whether methylation will lead to increased or decreased gene expression is complicated [79]. Although heavy methylation near transcription start sites canonically represses gene expression, hypomethylated regions, considered “transcriptionally permissive,” can be associated with either transcriptional activity or inactivity [80,81]. In addition, promoter methylation has now been found to bind certain transcription factors [82], can also promote transcription when it occurs at insulator binding sites [83], and can sometimes be overridden by gene body methylation [79].

Moreover, methylation in adulthood may reflect vestiges of transcriptional states in development rather than or in addition to present transcriptional states [84,85]. In particular, previous studies have demonstrated that methylation in adulthood at developmentally expressed genes can provide a window into past gene expression in neurons [85], which is of interest in itself, given challenges and limitations to assaying methylation directly during development in humans and nonhuman primates, especially great apes. Thus, while we believe the species- and brain structure-specific regDMRs in our analysis that exhibit corresponding changes in gene expression are of great interest, we also think the regDMRs unassociated with differences in adult gene expression warrant consideration and are nevertheless promising candidates for further study, particularly given many of their associations with genes involved in neurodevelopment.

Our study has a number of important limitations. First, our use of a microarray for humans biases our results toward conserved and human functional regions. We are thus likely missing divergent regions, such as enhancers specific to the other species. This may be why we observed overall less chimpanzee-specific methylation and less overlap in chimpanzee-specific methylation between brain structures. Also, we did not assay non-CpG methylation. Methylation at cytosines outside of CpG contexts has an important role in the brain [86]; species comparisons in non-CpG methylation are therefore an important part of the picture that is absent from our current analysis. It is further impossible to distinguish presently between evolved differences in developmental programs and species differences that may be environmental in origin. Indeed, given the known role of methylation in environmental programming [87] and the protracted postnatal development of the brain [88], it is highly likely that growing up human in our highly constructed environments contributes to some of the species differences we observe, as it likely does to the unique cognitive traits we are interested in understanding. Both evolved and environmentally or culturally induced differences are of interest for understanding human cognitive uniqueness, but have different bases that are not readily teased apart. Finally, heterogeneity in sample collection and processing could influence some of our findings. Although the limited availability of human and nonhuman primate tissues is a challenge that can result in suboptimal experimental variation like that in our study [77], the relative stability of methylation, for example, relative to RNA, under various storage conditions, or its “technical reliability” [79,89-92], makes it unlikely that batch effects resulting from differences in storage conditions are a major issue.

In conclusion, we identified human-specific patterns of methylation in the cerebellum, in addition to the DLPFC, which may reflect evolved and/or developmental neurobiological differences relevant to human-specific traits, like language, tool use, and cultural behavior, in these brain structures. Notably, we found greater human-specific differential methylation in the cerebellum. This finding may be driven by strong signal from excitatory neurons, which make up the bulk of cells in the cerebellum. Indeed, the cerebellum houses the majority of the brain’s neurons and plays a role in many important cognitive, as well as motor, processes [25,26,31,32]. Our results lend further weight to growing interest in cerebellar evolution, as well as provide crucial comparative context for interpreting studies of differential methylation in the cortex. In addition, we found that human-specific differential methylation overlapped genes in pathways associated with neurodevelopment, which may reflect aspects of human-specific developmental trajectories, and synaptic plasticity and lipid metabolism, which are thought to be altered in humans [66,68]. These differences may hold clues to the molecular underpinnings of human cognitive specializations. Moreover, some of these genes are also implicated in neuropathology and thus may offer clues to human disease susceptibilities. Few genes showing human-specific methylation patterns at putative regulatory elements showed human-specific expression; nevertheless, these genes include some interesting candidates for further study, such as CPNE6 (copine-6), which plays a role in learning and memory and shows human-specific hypomethylation and upregulation in the cerebellum. Our results highlight the value of tissue-specific species comparisons of methylation and are consistent with an important role for the cerebellum in human brain evolution.

Methods

Ethics statement

Brain specimens were obtained from the NIH NeuroBioBank, the National Chimpanzee Brain Resource, the Southwest National Primate Research Center, and the California National Primate Research Center with the approval of The George Washington University Institutional Animal Care and Use Committee (Protocol #A454). No living animals were used.

Study subjects

Specimens were collected postmortem from 22 individuals of three species (7 humans, 8 chimpanzees, and 7 rhesus macaques) (Fig 1A and S1 Table) who died from causes unrelated to this research. All human specimens were categorized as healthy controls. Brain specimens were frozen after postmortem intervals (PMIs) of less than 24 hours, except in the case of one of the human individuals (PMI: 26.7 hours). We selected a mix of relatively young and older adult subjects from each species. We obtained the specimens from the NIH NeuroBioBank, the National Chimpanzee Brain Resource, the Southwest National Primate Research Center, and the California National Primate Research Center. Following necropsy, brain specimens were coronally sectioned and slabs were stored at -80°C. Both brain regions came from the same individual, except in one case for humans, where the two regions came from different individuals of a similar age but different sex.

Tissue dissection

The dorsolateral prefrontal cortex (DLPFC, corresponding to Brodmann’s area 46) and lateral cerebellum (Crus I and Crus II; “cerebellum” hereafter) were dissected from frozen chimpanzee and macaque brain sections kept chilled on dry ice using published species-specific brain atlases [27,93,94] (Fig 1B). Dissected tissue was then transferred to RNALater (Ambion, Austin, TX, USA) preservation buffer and stored at -20°C in solution until DNA extraction. Human samples from the structures of interest were requested from the NIH NeuroBioBank and anatomical dissections followed by snap-freezing and pulverization were carried out by their staff. All tissue derived from the left hemisphere, except for in the case of two of the chimpanzees for which only the right hemisphere was available.

DNA extraction and microarray analysis

DNA was extracted using the DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany). Prior to extraction, RNALater was washed from macaque and chimpanzee samples with PBS buffer. We quantified DNA extracts on a Nanodrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and brought them to a concentration of ~70 ng/μl, either through dilution with PCR-grade water or concentration using Microcon-30kDa Centrifugal Filter Unit columns (Millipore Sigma, Burlington, MA, USA). DNA was then bisulfite-converted and genome-wide methylation levels were profiled by microarray (Illumina Infinium EPIC Methylation BeadChip; “EPIC array” hereafter) at the Yale Center for Genome Analysis. A mix of species, ages, and sexes were profiled on each chip to avoid batch effects.

Data preprocessing

We first filtered the raw intensity data to remove probes with spectral intensities not significantly different from background levels, that do not target CpG dinucleotides, that contain known SNPs, that are on the sex chromosomes, and for which fewer than 3 beads were counted for 5% or more of the samples with the ChAMP v2.18.3 [95] R package. We excluded one macaque cerebellum sample due to a high proportion of probes that did not show spectral intensities above background levels. We then normalized the data using Beta Mixture Quantile dilation (BMIQ), which accounts for the two different probe types on the EPIC array [96]. Because the Illumina methylation microarrays are designed to target human sites, we limited all analyses to CpG sites that should also be efficiently targeted in chimpanzees and rhesus macaques. Specifically, we restricted the dataset to CpG sites targeted by probes that map to the chimpanzee (panTro2.1.4) and rhesus macaque (Mmul8.0.1) genomes with 0–2 mismatch[s], without any mismatches within 5 bp of the target CpG, and without known SNPs in the target species following [12].

Brain tissue is a composite of multiple cell types. We used surrogate variable analysis, a robust, reference-free approach that can account for cell type heterogeneity [80,97]. Specifically, for both our DLPFC and cerebellum data separately, we first estimated the number of cell types (k) in the data using a random matrix theory approach to identify the number of variance components in the data above the number expected under a Gaussian Orthogonal Ensemble (GOE) using the EstDimRMT function in R package isva v1.9 [98]. EstDimRMT returned an optimal k of two variance components/cell types for each tissue. We then estimated two surrogate variables using a re-weighted least squares approach [99] with the sva function in the R package sva v3.36.0 [100]. We used singular value decomposition to assess remaining batch effects [101] in both the DLPFC and cerebellum datasets with the champ.SVD function in the ChAMP package v2.18.3 [95] and found no significant variance associated with array or slide.

We identified differential methylation in two ways. We first identified differential methylation between humans and the two nonhuman primates, chimpanzees and macaques, in a dataset of CpGs present in all three species. These datasets consisted of 165,398 and 165,740 probes for the DLPFC and cerebellum, respectively. The sparsity of CpGs in the dataset including macaques due to decreased sequence conservation with phylogenetic distance, however, limits our power to discover stretches of differential methylation in all three species. We thus additionally took a second, two-step approach to identifying species-specific methylation patterns in which we identified differential methylation between humans and chimpanzees only. After filtering, 556,428 and 557,796 probes remained in our chimpanzee-human DLPFC and cerebellum datasets, respectively. We then used overlapping, sparser macaque data to polarize these differences.

In both approaches, we first identified CpG sites showing differential methylation (differentially methylated positions, or DMPs) among species in each brain structure using the lmFit linear regression function in the limma v3.44.3 R package with eBayes coefficient shrinkage [102]. We included the covariates of age class, sex, and hemisphere, as well as the two surrogate variables to account for cell type heterogeneity (S1 Table). We considered DMPs significant at a false discovery rate [103] of 5%. We further required that DMPs exhibit a mean beta difference between humans and nonhuman primates of at least 0.15.

We next identified differentially methylated regions (DMRs)—or contiguous stretches of the genome enriched for differential methylation—among species using the DMRcate v2.2.3 R package, which calls DMRs based on Gaussian kernal smoothed methylation levels across stretches of the genome [104]. We required DMRs to contain at least five CpGs and that at least three of these CpGs be DMPs, as determined in our previous analysis. For the human-chimpanzee DMRs, we then determined whether human-chimpanzee differential methylation was human- or chimpanzee-specific by polarizing our comparisons using the macaque data. Specifically, we determined which human-chimpanzee DMRs overlapped at least three macaque-human or macaque-chimpanzee DMPs identified following the procedure described above and whether these DMPs represented at least 50% of the CpGs comprising the DMR or covered at least 50% of the DMR’s length. The resulting DMRs were determined to be human- or chimpanzee-specific DMRs.

We determined the degree to which DMRs in each structure were specific to that structure by assessing the overlap between the two sets of brain structure DMRs, as well as their correspondence with DMRs in blood, using a human and chimpanzee dataset. Blood methylation data for chimpanzees comprised a dataset we previously generated for 87 chimpanzee blood samples from 73 individuals using the EPIC array [105]. Human blood methylation data generated on the 450K array was downloaded for 274 individuals [106-108] from the Gene Expression Omnibus online database (accessions: GSE40279, GSE87571, GSE56105). Only probes on both the EPIC and 450K arrays that are expected to effectively target CpG sites also present in our brain methylation dataset were retained in this analysis. The final dataset was comprised of 129,003 probes. Data were quality filtered and normalized, surrogate variables estimated to account for cell type heterogeneity, and DMPs and DMRs identified using the same procedure as for the brain data. We then determined overlap between blood and brain DMRs, as well as between the two brain structures. DMRs were considered to be overlapping if they shared more than half of their constituent CpGs or half of their genomic range with a DMR in another tissue. DMRs were considered to be structure- or tissue-specific if they did not show overlap with a DMR in the other brain structure or blood.

We further determined which genomic elements DMRs spanned and which genes, if any, they were associated with by annotating CpGs within DMRs using the Illumina EPIC array annotation in the ChAMP v2.18.3 R package [95]. We considered DMRs comprised by DNA within 1500 bp of transcription start sites or covering untranslated regions or the first exon to be potentially regulatory associated (regDMRs). DMRs covering gene bodies or in intergenic regions were not considered to be regDMRs.

Correspondence with gene expression

We assessed correspondence with gene expression of human and chimpanzee brain structure-specific regDMRs using previously published data for the two structures for the three species [109] (http://evolution.psychencode.org/#). We downloaded count data and analyzed it using the DESeq2 v1.22.2 R package [110]. We removed genes with fewer than ten counts and used the DESeq function to correct for differences in count due to library size, estimate data dispersion, and fit a differential expression model for each pair of species. We then identified differentially expressed genes in humans as genes that were significantly differentially expressed (adjusted p-value < 0.05) and exhibited a log fold change of at least 1.5 in both the human-chimpanzee and human-macaque comparison. We also identified genes that were differentially expressed in chimpanzees using an equivalent procedure. We also examined whether genes that show uniquely human expression patterns and are associated with human-specific regDMRs fall in the same pathways. Specifically, we determined whether any genes annotated with the biological process GO term “chemical synaptic transmission” (GO:0007268)–a term that annotated two genes (DLGAP1, CPNE6) associated with human-specific regDMRs in the cerebellum–showed uniquely human expression patterns in the cerebellum. We retrieved a list of genes annotated with this term using the get_anno_genes function in the R package GOfuncR v1.8.0 [111] and assessed whether these genes showed a uniquely human pattern of expression.

Correspondence with previous studies

We assessed the overall correspondence of our DLPFC data with previously published prefrontal cortex WGBS data for humans and chimpanzees [16]. We calculated the mean methylation level for each species at each CpG site in each dataset. The chimpanzee WGBS data was aligned to the panTro2.0 genome build, so we used the liftover function in the R package rtracklayer v1.48.0 [112] to identify the corresponding hg19 coordinates. We then assessed the correlation between average methylation for each species in the two datasets across all sites overlapping between the two datasets. We used to same approach to measure correspondence of the prefrontal cortex WGBS data with our cerebellum data to assess region specificity.

Human-specific regDMRs in both brain structures.

A = PRSS50 and B = EGFLAM. Lines are loess smoothed methylation values for each species across the DMR and each point represents raw methylation values for each individual at each CpG site within the DMR ranges. TSS200 = within 200 bp of a transcription start site and TSS1500 within 1500 bp of a transcription start site. Chromosomal coordinates are in base pairs and refer to human genome build hg19.

(TIF)

Click here for additional data file.

Human-specific cross-tissue regDMRs.

A = C1orf65 and B = BMPER. Lines are loess smoothed methylation values for each species across the DMR and each point represents raw methylation values for each individual at each CpG site within the DMR ranges. TSS200 = within 200 bp of a transcription start site and TSS1500 within 1500 bp of a transcription start site. Chromosomal coordinates are in base pairs and refer to human genome build hg19.

(TIF)

Click here for additional data file.

Cerebellum-specific chimpanzee regDMRs associated with STYL1 and STYL1 gene expression.

A = methylation, B = gene expression. Lines are loess smoothed methylation values for each species across the DMR and each point represents raw methylation values for each individual at each CpG site within the DMR ranges. TSS200 = within 200 bp of a transcription start site and TSS1500 within 1500 bp of a transcription start site. Chromosomal coordinates are in base pairs and refer to human genome build hg19.

(TIF)

Click here for additional data file.

Chimpanzee-specific DMR overlapping DRD2 in the cerebellum.

Lines are loess smoothed methylation values for each species across the DMR and each point represents raw methylation values for each individual at each CpG site within the DMR ranges. Chromosomal coordinates are in base pairs and refer to human genome build hg19.

(TIF)

Click here for additional data file.

Study subjects.

(XLSX)

Click here for additional data file.

Species and region-specific DMRs.

(XLSX)

Click here for additional data file.

Genes showing uniquely human expression in the cerebellum annotated with the GO term chemical synaptic transmission.

(XLSX)

Click here for additional data file.

Human and chimpanzee brain region-specific regDMRs annotated with GO terms.

(XLSX)

Click here for additional data file.

11 Sep 2020

Dear Dr Guevara,

Thank you very much for submitting your Research Article entitled 'Comparison with chimpanzees and rhesus macaques reveals distinctive epigenetic features of the human prefrontal cortex and cerebellum' to PLOS Genetics. Your manuscript was fully evaluated at the editorial level and by independent peer reviewers. The reviewers appreciated the attention to an important problem, but raised some substantial concerns about the current manuscript. Based on the reviews, we will not be able to accept this version of the manuscript, but we would be willing to review again a much-revised version. We cannot, of course, promise publication at that time.

Should you decide to revise the manuscript for further consideration here, your revisions should address the specific points made by each reviewer. We will also require a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript.

If you decide to revise the manuscript for further consideration at PLOS Genetics, please aim to resubmit within the next 60 days, unless it will take extra time to address the concerns of the reviewers, in which case we would appreciate an expected resubmission date by email to plosgenetics@plos.org.

If present, accompanying reviewer attachments are included with this email; please notify the journal office if any appear to be missing. They will also be available for download from the link below. You can use this link to log into the system when you are ready to submit a revised version, having first consulted our Submission Checklist.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see our guidelines.

Please be aware that our data availability policy requires that all numerical data underlying graphs or summary statistics are included with the submission, and you will need to provide this upon resubmission if not already present. In addition, we do not permit the inclusion of phrases such as "data not shown" or "unpublished results" in manuscripts. All points should be backed up by data provided with the submission.

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool.  PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

PLOS has incorporated Similarity Check, powered by iThenticate, into its journal-wide submission system in order to screen submitted content for originality before publication. Each PLOS journal undertakes screening on a proportion of submitted articles. You will be contacted if needed following the screening process.

To resubmit, use the link below and 'Revise Submission' in the 'Submissions Needing Revision' folder.

[LINK]

We are sorry that we cannot be more positive about your manuscript at this stage. Please do not hesitate to contact us if you have any concerns or questions.

Yours sincerely,

Takashi Gojobori

Associate Editor

PLOS Genetics

Bret Payseur

Section Editor: Evolution

PLOS Genetics

Editor's comments:

In this paper, the authors studied epigenetics related to the evolution of human neural specialization, measuring genome-wide DNA methylation levels in brain tissues from humans, chimpanzees, and macaques. Their findings of evolutionary changes in species-specific methylation with its implication to neurobiological function are significant and interesting.  All the three reviewers agreed with this point.

However, Reviewers 2 and 3 questioned the authors’ usage of the “EPIC microarray“ rather than whole-genome bisulfite sequencing for a less biased data set.  In particular, Reviewer 2 suggested rejection of this manuscript, pointing out that the choice to use EPIC array to investigate the brain epigenomic and the absence of any cell (nuclei) type enrichment significantly reduced the potential impact of this study.  Reviewer 3 also stated that the major methodological issue that one see with this study is profiling methylation sites by microarray rather than whole-genome bisulfite sequencing for a less biased data set.

Therefore, I would like to see the authors’ response to the above-mentioned points regarding the methodology.

Moreover, I have to agree with the following particular points of Reviewers 1 and 3:

As Reviewer 1 pointed out that the statistical analyses employed need to be more detailed, and potential confounding batch effects should be further considered. This Reviewer further continued that additionally, as the authors note, research on DNA methylation in brain tissues from primates has been previously examined, so the originality of this study should be more explicitly stated and reinforced.

Reviewer 3 also stated that a more focused approach would have served this study better: Identify the areas of greatest difference between species, narrow down on the most interesting or understudied pathways, and experimentally test whether hypo- or hypermethylation correlates with increased or decreased expression of the genes in question to construct narrow hypotheses around these specific data.

Those comments should be taken into account seriously by the authors when they make revision of the manuscript.

In summary, I would request the authors to make major revision and send it back to us.  Then, the revised manuscript will be examined by the reviewers again, for making final decision.

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: Review is attached.

Reviewer #2: The study by Guevara et al, compared DNA methylation pattern for two brain regions: DLPFC, cerebellum across human, chimpanzees, and an old-world monkey rhesus macaques. This was a well-designed study to investigate the evolution of the epigenome and gene regulation in primates. The study included an impressive set of primary brain tissue specimens, with sample quality control and dissection carefully performed. In addition, DLPFC and cerebellum samples were collected from paired individuals, increasing the consistency of the dataset, and potentially allow the controlling of individual sample difference with regression models. However, the choice to use EPIC array to investigate the brain epigenomic and the absence of any cell (nuclei) type enrichment significantly reduced the potential impact of this study.

1. A number of groups (e.g. Guo 2013 Nat Neuro, Lister 2013 Science, Gabel 2015 Nature) have reported the presence of a large quantity of non-CG methylation in mammalian brains, which appears to be a very unique characteristic of the brain compared to other tissues. Likely non-CG methylation is relevant to the evolution of the brain given its uniquely high level of accumulation in the brain. Given these findings, the authors should consider using whole-genome bisulfite sequencing that can more effectively measure non-CG methylation. If cost is a concern here, what about EPIC capture sequencing assay?

2. It is well known that brain cell types have very diverse methylation profiles, especially between neuronal cells and glial cells. It is fully feasible to separate neuronal and non-neuronal cells using anti-NeuN antibody. Kozlenkov et al., 2015 NAR has also developed a method to separate excitatory and inhibitory neuronal nuclei using antibodies against transcriptional factor proteins. These methods have been tested with primary human brain tissue and should be readily applicable to primate brain tissues.

3. Separating neurons and glia would be important not just for extracting cell type specific signature, but also to ensure the inter-species difference was not due to the variable ratio between neuronal and non-neuronal cells. Differences in cell-type composition is a major confounding factor for epigenetic epidemiology studies and may have an impact on inter-species comparison too.

4. The manuscript should include some technical description of the specimens, the method for methylation profiling and a brief discussion of the statistical method for DMR calling at the beginning of the result section. It was difficult to put the scale of DMR numbers in the context with no technical description at all in the main text.

Reviewer #3: The manuscript by Guevara et al. presents a relevant contribution to species-and tissue-related differences in global methylation of the dorsolateral prefrontal cortex and lateral cerebellum, brain areas considered of specific importance in the evolution of human cognition. Identified differentially methylated regions included genes that encode factors relevant to energy metabolism, neuroinflammation, neural degeneration, and neurodevelopment. Diminishing the impact of these discoveries, human-specific differentially methylated regions showed extensive overlap with those identified in blood, suggesting that some epigenetic marks may be common to nonneural cell types. Further diminishing the impact of this work, only modest correspondence between human-specific differential methylation and differential gene expression was observed. While the overall scope of work appears appropriate and some new insight may have been gained here, overall advancement for the field appears modest and somewhat dampens enthusiasm for this paper. It appears to me that a more focused approach would have served this study better: Identify the areas of greatest difference between species, narrow down on the most interesting or understudied pathways, and experimentally test whether hypo- or hypermethylation correlates with increased or decreased expression of the genes in question to construct narrow hypotheses around these specific data. Please see below a more detailed critique.

The major methodological issue that I see with this study is profiling methylation sites by microarray rather than whole-genome bisulfite sequencing for a less biased data set. The authors explain the problem themselves “Because the Illumina methylation microarrays are designed to target human sites, we limited all analyses to CpG sites that should also be efficiently targeted in chimpanzees and rhesus macaques.” The bias that this approach introduces cannot possibly overstated and may stood in the way of discovering overlooked aspects of human-specific methylation and gene regulation.

The experimental design remains somewhat obscure and I recommend the authors provide additional clarifications. I assume that the numbers of DMPs and DMRs provided inn Fig. 1 and Table 1 are in relation to the reference methylome of macaque, but nowhere is this clearly explained. Please clarify.

The authors discuss multiple very interesting pathways involving genes that associate with DMRs, such as glucose metabolism, neuroinflammation etc., but do not provide data on whether the genes in question are upregulated or downregulated in these cases, which seems to me the most relevant question.

I recommend that whenever in the results the number of overlapping DMRs is provided, to also indicate what percentage of the total relevant DMRs this number corresponds to.

Results of figures 3 and 4 are only mentioned in the discussion and there only partially. Please present these data in the results section as well.

In the discussion the authors state: “Many human-specific DMRs also overlapped genes involved in neurodevelopment and neurodevelopmental disorders, especially autism spectrum disorder (ANKS1B, GCNT2, DIP2C, RAI1, MYOZ3, DIP2C, DLGAP1, ZNF8) and schizophrenia (NT5DC2, SNED1, RAI1, MYOZ3, FAM193A).” How do the listed genes associate with the ID and ASD modules established by Parikshak et al. 2013? Are prenatal or postnatal modules of greater relevance?

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Genetics data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

Submitted filename: PGENETICS-D-20-01121_review.pdf

Click here for additional data file.

11 Jan 2021

Submitted filename: PLoS_Genetics_response_to_reviewer_comments_2021.01.09.docx

Click here for additional data file.

17 Feb 2021

Dear Dr Guevara,

Thank you very much for submitting your Research Article entitled 'Comparative analysis reveals distinctive epigenetic features of the human cerebellum' to PLOS Genetics.

The manuscript was fully evaluated at the editorial level and by independent peer reviewers. The reviewers appreciated the attention to an important topic but identified some concerns that we ask you address in a revised manuscript

We therefore ask you to modify the manuscript according to the review recommendations. Your revisions should address the specific points made by each reviewer.

In addition we ask that you:

1) Provide a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript.

2) Upload a Striking Image with a corresponding caption to accompany your manuscript if one is available (either a new image or an existing one from within your manuscript). If this image is judged to be suitable, it may be featured on our website. Images should ideally be high resolution, eye-catching, single panel square images. For examples, please browse our archive. If your image is from someone other than yourself, please ensure that the artist has read and agreed to the terms and conditions of the Creative Commons Attribution License. Note: we cannot publish copyrighted images.

We hope to receive your revised manuscript within the next 30 days. If you anticipate any delay in its return, we would ask you to let us know the expected resubmission date by email to plosgenetics@plos.org.

If present, accompanying reviewer attachments should be included with this email; please notify the journal office if any appear to be missing. They will also be available for download from the link below. You can use this link to log into the system when you are ready to submit a revised version, having first consulted our Submission Checklist.

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

Please be aware that our data availability policy requires that all numerical data underlying graphs or summary statistics are included with the submission, and you will need to provide this upon resubmission if not already present. In addition, we do not permit the inclusion of phrases such as "data not shown" or "unpublished results" in manuscripts. All points should be backed up by data provided with the submission.

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript.

PLOS has incorporated Similarity Check, powered by iThenticate, into its journal-wide submission system in order to screen submitted content for originality before publication. Each PLOS journal undertakes screening on a proportion of submitted articles. You will be contacted if needed following the screening process.

To resubmit, you will need to go to the link below and 'Revise Submission' in the 'Submissions Needing Revision' folder.

[LINK]

Please let us know if you have any questions while making these revisions.

Yours sincerely,

Takashi Gojobori

Associate Editor

PLOS Genetics

Bret Payseur

Section Editor: Evolution

PLOS Genetics

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: The authors have satisfactorily addressed all of my prior comments. In particular, the refocusing of the paper to emphasize the cerebellum very much clarifies and enhances the originality of this study. The additional details provided in the methods are also an excellent improvement. Finally, while the authors do a nice job of highlighting their results and the insights gained from them, they are also clear about the limitations of this work.

Minor Comments

[line 472] What does “age class” mean? If this is something different than age in years, please specify this variable in Table S1.

Some minor typos are present throughout the paper and should be corrected. For example:

• [line 108] “languages areas” should be changed to “language areas”

• [line 113] remove the comma between “molecular” and “studies”

• [line 119] add a comma after in-text citation “[27-29]”

Reviewer #3: The substantially revised manuscript by Guevara et al. presents an partial improvement over the initial submission addressing some points raised by the reviewers, particularly regarding methodological detail. However, some major points were not addressed, including my earlier suggestion to provide a greater focus to this study by identifying the areas of greatest difference between species, narrow down on the most interesting or understudied pathways, and experimentally test whether hypo- or hypermethylation correlates with increased or decreased expression of the genes in question to construct narrow hypothesis around these specific data. While I am very sympathetic to the argument about current limitations placed on most of us, even the existing data without additional experiments may have sufficed to meet part of this suggestion. For instance, I find Table 1 dissatisfying by not listing genes according to relevant gene ontology domains using common tools that may further highlight pathways of relevance.

In addition, the paper’s focus has been shifted by emphasizing the cerebellum as a structure of potential significance to human evolution, but throughout the manuscript and particularly so in the discussion this position is continuously weakened or the evidence supporting a disproportionate significance of the human cerebellum not sufficiently highlighted.

The results section has been shortened to a degree that makes it difficult to follow the process and rationale guiding experiments. Providing some additional detail, currently dispersed in introduction and methods should help. For instance, the results section is initiated by directly presenting the methylation data with no mention of the process leading to these data. While the introduction and methods sections provide more details, some of it has to make it into the results section as well. Specimen selection, amount of tissue collected, lysis procedure, and reference to the chip used for analysis are some essentials to provide context. In the process, please also include a reference to Fig 1 in the results section.

Line 165: Many (~7%) of DMRs identified were evident in both brain structures. I don’t think that 7% qualifies as many. I think “some” or a “minority” seems more appropriate.

Line 169: Overall, correspondence across tissues was high, with average global DLPFC and cerebellum methylation showing a correlation of 0.91-0.92 in all three species.

Line 208: Substantial species divergence in methylation of the cerebellum is notable giving a long standing focus on the prefrontal cortex in comparative neuroscience. Our finding affirms the importance of the cerebellum in human brain evolution.

The above two statements from results and discussion respectively disagree in my opinion. At .91-.92 correlation of both DLPFC and cerebellum between species, both structures show comparable divergence. Also, I am not sure if this level of between species divergence can be considered substantial, as it corresponds well to that seen in blood too.

**********

Have all data underlying the figures and results presented in the manuscript been provided?

Large-scale datasets should be made available via a public repository as described in the PLOS Genetics data availability policy, and numerical data that underlies graphs or summary statistics should be provided in spreadsheet form as supporting information.

Reviewer #1: Yes

Reviewer #3: Yes

**********

PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #3: No

6 Mar 2021

Submitted filename: PLoS_Genetics_response_to_reviewer_comments_2021.03.04.docx

Click here for additional data file.

24 Mar 2021

Dear Dr Guevara,

We are pleased to inform you that your manuscript entitled "Comparative analysis reveals distinctive epigenetic features of the human cerebellum" has been editorially accepted for publication in PLOS Genetics. Congratulations!

Before your submission can be formally accepted and sent to production you will need to complete our formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Please note: the accept date on your published article will reflect the date of this provisional acceptance, but your manuscript will not be scheduled for publication until the required changes have been made.

Once your paper is formally accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you’ve already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosgenetics@plos.org.

In the meantime, please log into Editorial Manager at https://www.editorialmanager.com/pgenetics/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production and billing process. Note that PLOS requires an ORCID iD for all corresponding authors. Therefore, please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field.  This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager.

If you have a press-related query, or would like to know about making your underlying data available (as you will be aware, this is required for publication), please see the end of this email. If your institution or institutions have a press office, please notify them about your upcoming article at this point, to enable them to help maximise its impact. Inform journal staff as soon as possible if you are preparing a press release for your article and need a publication date.

Thank you again for supporting open-access publishing; we are looking forward to publishing your work in PLOS Genetics!

Yours sincerely,

Takashi Gojobori

Associate Editor

PLOS Genetics

Bret Payseur

Section Editor: Evolution

PLOS Genetics

www.plosgenetics.org

Twitter: @PLOSGenetics

----------------------------------------------------

Comments from the reviewers (if applicable):

----------------------------------------------------

Data Deposition

If you have submitted a Research Article or Front Matter that has associated data that are not suitable for deposition in a subject-specific public repository (such as GenBank or ArrayExpress), one way to make that data available is to deposit it in the Dryad Digital Repository. As you may recall, we ask all authors to agree to make data available; this is one way to achieve that. A full list of recommended repositories can be found on our website.

The following link will take you to the Dryad record for your article, so you won't have to re‐enter its bibliographic information, and can upload your files directly:

http://datadryad.org/submit?journalID=pgenetics&manu=PGENETICS-D-20-01121R2

More information about depositing data in Dryad is available at http://www.datadryad.org/depositing. If you experience any difficulties in submitting your data, please contact help@datadryad.org for support.

Additionally, please be aware that our data availability policy requires that all numerical data underlying display items are included with the submission, and you will need to provide this before we can formally accept your manuscript, if not already present.

----------------------------------------------------

Press Queries

If you or your institution will be preparing press materials for this manuscript, or if you need to know your paper's publication date for media purposes, please inform the journal staff as soon as possible so that your submission can be scheduled accordingly. Your manuscript will remain under a strict press embargo until the publication date and time. This means an early version of your manuscript will not be published ahead of your final version. PLOS Genetics may also choose to issue a press release for your article. If there's anything the journal should know or you'd like more information, please get in touch via plosgenetics@plos.org.

13 Apr 2021

PGENETICS-D-20-01121R2

Comparative analysis reveals distinctive epigenetic features of the human cerebellum

Dear Dr Guevara,

We are pleased to inform you that your manuscript entitled "Comparative analysis reveals distinctive epigenetic features of the human cerebellum" has been formally accepted for publication in PLOS Genetics! Your manuscript is now with our production department and you will be notified of the publication date in due course.

The corresponding author will soon be receiving a typeset proof for review, to ensure errors have not been introduced during production. Please review the PDF proof of your manuscript carefully, as this is the last chance to correct any errors. Please note that major changes, or those which affect the scientific understanding of the work, will likely cause delays to the publication date of your manuscript.

Soon after your final files are uploaded, unless you have opted out or your manuscript is a front-matter piece, the early version of your manuscript will be published online. The date of the early version will be your article's publication date. The final article will be published to the same URL, and all versions of the paper will be accessible to readers.

Thank you again for supporting PLOS Genetics and open-access publishing. We are looking forward to publishing your work!

With kind regards,

Katalin Szabo

PLOS Genetics

On behalf of:

The PLOS Genetics Team

Carlyle House, Carlyle Road, Cambridge CB4 3DN | United Kingdom

plosgenetics@plos.org | +44 (0) 1223-442823

plosgenetics.org | Twitter: @PLOSGenetics