Human-specific CpG ‘beacons’ identify human-specific prefrontal cortex H3K4me3 chromatin peaks

Figures & data

Figure 1.  UCSC browser display showing the colocation of overlapping human-specific H3K4me3 peak (top level mauve) and extreme CpG ‘beacon’ clusters (red) at the 5′ CHL1 locus.

All figures are viewable in color online at www.tandfonline.com/doi/full/10.2217/epi.13.74

Figure 2.  Human chromosomes 2, 3, 16 and 20.

Outer ring: Chromosomes with banding patterns. Dense edge ring: Enredo-Pecan-Ortheus [10,11] h1c1o1 coverage of these chromosomes. Blue peaks: extreme CpG ‘beacon’ peak locations. Human-specific H3K4me3 location with false discovery rate log p values compared to chimpanzee (orange), macaque (green) and overlapping chimpanzee and macaque (olive). Location of ANKRD11, CHL1, DPP10, HAR1A and PSD4 indicated with colocating CpG ‘beacon’ clusters and human-specific H3K4me3. The highlighted yellow wedge is the location of the chromosome 2qFus (2q13–2q14.1) with enrichment of both extreme CpG ‘beacon’ clusters and of Human-specific H3K4me3 peaks (χ²: p = 1.045 × 10^-5).

Figure 3.  Stronger human-specific H3K4m3 peaks identified (FDR p value) with CpG ‘beacon’ density in both human versus chimpanzee (A), and versus rhesus macaque (B) (Wilcoxon test CpG ‘beacon’ 0 vs > = 20, p = 3.57 × 10^-3 chimpanzee and p = 2.50 × 10^-2 macaque (p = 0 values capped at 10^-200).

This analysis is within the primate-alignable (EPO [10,11]) portion of the human genome. EPO: Enredo-Pecan-Ortheus; FDR: False discovery rate.

Figure 4.  Comparison of studies: proportion of identified loci with overlapping human-specific H3K4me3.

CpG ‘beacon’ clusters identify hs-H3K4me3 peaks more successfully than extreme human nucleotide change or studies of selection. Scatter plot of permutated data for random selection of sets of one to 1000 loci from the total of 18,665 human peaks (1000× randomization each) that then overlap the 410 hs-H3K3me4 loci. Upper line: Loess regression line of maximum simulation value; golden points: average for each random selection with average Loess regression line overlaid in grey; red squares: CpG ‘beacon’ clusters; green squares: CpG ‘beacon’ clusters adjusted for the fact that the primate EPO alignment only contains 257 hs-H3K4me3 peaks; blue squares: selective studies – HACNS (HACNS – 0/992) [43] and ANC (ANC – 6/1356) [44], Select_2x, _3x and _4+x are overlaps with studies combined by Akey [42] with loci colocating in 2, 3 or 4 or more studies. HARs are the 49 top HARs [21]. HAR_202 is an expanded set of 202 HARs, with this set the only nonbeacon analysis to perform well, as this larger grouping has a strong overlap with CpG ‘beacons’ (∼85.7% of the hs-H3K4me3 overlapping CpG ‘beacons’ ≥10/kb). ANC: Accelerated noncoding; CB_10: CpG ‘beacon’ clusters ≥10/kb; CB_15: CpG ‘beacon’ clusters ≥15/kb; CB_20: CpG ‘beacon’ clusters ≥20/kb; HAR: Human accelerated regions; HACNS: Human accelerated conserved noncoding; hs-H3K4me3: Human-specific H3K4me3.

Figure 5.  The hypothesis associating potential recombination-driven permissive chromatin formation with CPGi formation via CpG ’beacons’.

The histone methylase PRDM9 directs recombination via motif-specificity, with resultant biased gene conversion within these regions increasing GC content and therefore CpG numbers. Dense CpG regions are able to recruit chromatin-modifying factors such as Cpf1, which lead to the formation of permissive H3K4me3 chromatin domains. Due to the strong association of PRDM9 with recombination [51], ancestral alleles of this rapidly evolving DNA-binding factor [52] are hypothesized to be implicated in this recombination-associated biased gene conversion [39].

Table 1.  Human-specific CpG ‘beacon’ clusters with Human-specific H3K4me3.

CpG ‘beacons’/1 kb	Locus	Gene	Location	Distance to TSS	FDR Human-specific vChimpanzee	FDR Human-specific vMacaque	Supporting PFC H3K4me3^†	#OMIM
32	chr20:61732970-61734710	HAR1B	5′ promoter	0	5.302 × 10^-10	0.00 × 10^+00	Yes	*610556/*610557
26	chr2:113914422-113915745	PSD4	Upstream	15,814	4.101 × 10^-9	4.670 × 10^-11	Yes	*614442
22	chr2:115419442-115420562	DPP10/LOC389023	Intragenic	200,263	1.340 × 10^-38	0.00 × 10^+00	Yes	*608209
21	chr3:237734-240480	CHL1	5′ promoter	0	2.296 × 10^-42	3.501 × 10^-92	Yes	*607416/#613792
21	chr16:89322850-89324461	ANKRD11/ZNF778	3′	38,740	1.239 × 10^-27	0.00 × 10^+00	Yes	*611192/#148050
17	chr15:101,457,966-101,461,331	LRRK1	5′ promoter	0	4.105 × 10^-9	3.824 × 10^-47	Yes	*610986
17	chr3:2139259-2142710	CNTN4	5′ promoter	0	2.494 × 10^-25	1.156 × 10^-37	Yes	*607280/#613792
17	chr2:115918003-115921686	DPP10	5′ promoter	0	7.824 × 10^-29	1.046 × 10^-13	Yes	*608209

Human-specific CpG ‘beacon’ clusters ≥17/kb (empirical p < 0.05) with supporting human-specific H3K4me3 from Shulha et al. [24]. ^†Supporting human PFC H3K4me3 signature from the Maunakea et al. data [27]. FDR: False discovery rate; PFC: Prefrontal cortex; TSS: Transcription start site.

Paten B , HerreroJ , BealK , FitzgeraldS , BirneyE . Enredo and Pecan: genome-wide mammalian consistency-based multiple alignment with paralogs . Genome Res.   18 ( 11 ), 1814 – 1828 ( 2008 ).

 PubMed Web of Science ®Google Scholar

Paten B , HerreroJ , FitzgeraldS et al. Genome-wide nucleotide-level mammalian ancestor reconstruction . Genome Res.   18 ( 11 ), 1829 – 1843 ( 2008 ).

 PubMed Web of Science ®Google Scholar

Prabhakar S , ViselA , AkiyamaJA et al. Human-specific gain of function in a developmental enhancer . Science   321 ( 5894 ), 1346 – 1350 ( 2008 ).

 PubMed Web of Science ®Google Scholar

Bird CP , StrangerBE , LiuM et al. Fast-evolving noncoding sequences in the human genome . Genome Biol.   8 ( 6 ), R118 ( 2007 ).

 PubMed Web of Science ®Google Scholar

Akey JM . Constructing genomic maps of positive selection in humans: where do we go from here?   Genome Res.   19 ( 5 ), 711 – 722 ( 2009 ).

 PubMed Web of Science ®Google Scholar

Pollard KS , SalamaSR , LambertN et al. An RNA gene expressed during cortical development evolved rapidly in humans . Nature   443 ( 7108 ), 167 – 172 ( 2006 ).

 PubMed Web of Science ®Google Scholar

Baudat F , BuardJ , GreyC et al. PRDM9 is a major determinant of meiotic recombination hotspots in humans and mice . Science   327 ( 5967 ), 836 – 840 ( 2010 ).

 PubMed Web of Science ®Google Scholar

Oliver PL , GoodstadtL , BayesJJ et al. Accelerated evolution of the Prdm9 speciation gene across diverse metazoan taxa . PLoS Genet.   5 ( 12 ), e1000753 ( 2009 ).

 PubMed Web of Science ®Google Scholar

Capra JA , HubiszMJ , KostkaD , PollardKS , SiepelA . A model-based analysis of GC-biased gene conversion in the human and chimpanzee genomes . PLoS Genet.   9 ( 8 ), e1003684 ( 2013 ).

 PubMed Web of Science ®Google Scholar

Shulha HP , CrisciJL , ReshetovD et al. Human-specific histone methylation signatures at transcription start sites in prefrontal neurons . PLoS Biol.   10 ( 11 ), e1001427 ( 2012 ).

 PubMed Web of Science ®Google Scholar

Maunakea AK , NagarajanRP , BilenkyM et al. Conserved role of intragenic DNA methylation in regulating alternative promoters . Nature   466 ( 7303 ), 253 – 257 ( 2010 ).

 PubMed Web of Science ®Google Scholar