An Enhancer's Length and Composition Are Shaped by Its Regulatory Task

“Transcription factor (TF) binding sites were computationally predicted using Patser (Hertz and Stormo, 1999) with the position weight matrices (PWMs) from FlyFactor Survey (Zhu et al., 2011).”

“Though ChIP can experimentally identify TF-bound regions, existing data sets in the Drosophila embryo are low resolution, with ∼100 base pair peaks (Li et al., 2008; MacArthur et al., 2009; Roy et al., 2010), which are longer than the∼10 bp TF binding sites (Zhu et al.”

“To measure information in a motif, we calculated the Kullback-Leibler distance (Schneider et al., 1986; Stormo and Fields, 1998) between the motif and the composition of the intergenic regions of the genome I = L ∑ i = 1 ∑ b∈{A,C,G,T} pi ( b ) log2 pi ( b ) q ( b ) where L is the length of the…”

“Enhancers can be ∼10–1,000 bps long, with a couple to tens of TF binding sites (Blackwood and Kadonaga, 1998; Yáñez-Cuna et al., 2013).”

“…complex than the DV axis, with the AP axis consisting of 14 parasegments (Nasiadka et al., 2002) and the DV axis consisting of six germ layers and sublayers (Levine and Davidson, 2005), and therefore the patterning of the AP axis requires enhancers that drive more unique gene expression patterns.”

“Here, we analyzed the binding of TFs that were concurrently expressed with active enhancers based on the Berkeley Drosophila Genome Project in situ annotations (Tomancak et al., 2002, 2007; Hammonds et al., 2013).”

“…enhancer can only achieve distinguishability if it consists of a cluster of TF binding sites within a short distance (Wasserman and Fickett, 1998; Berman et al., 2002; Frith et al., 2002; Halfon et al., 2002; Markstein et al., 2002; Rebeiz et al., 2002; Li et al., 2007; Wunderlich and Mirny,…”

“Though ChIP can experimentally identify TF-bound regions, existing data sets in the Drosophila embryo are low resolution, with ∼100 base pair peaks (Li et al., 2008; MacArthur et al., 2009; Roy et al., 2010), which are longer than the∼10 bp TF binding sites (Zhu et al.”

“Here, we analyzed the binding of TFs that were concurrently expressed with active enhancers based on the Berkeley Drosophila Genome Project in situ annotations (Tomancak et al., 2002, 2007; Hammonds et al., 2013).”

“Of these enhancers, we consider the 3,580 enhancer candidates that were successfully refined to the putative minimal enhancers using functional genomics (Kvon et al., 2014).”

“For example, autoregulatory enhancers like ftz_up, ftz_zebra, and gt_minus1 (Hiromi et al., 1985; Hiromi and Gehring, 1987; Hoermann et al., 2016) have lower task complexity because they reinforce the expression patterns determined by other enhancers, and therefore, may not be consistent…”

“…litter the genome, and an enhancer can only achieve distinguishability if it consists of a cluster of TF binding sites within a short distance (Wasserman and Fickett, 1998; Berman et al., 2002; Frith et al., 2002; Halfon et al., 2002; Markstein et al., 2002; Rebeiz et al., 2002; Li et al.,…”

“For example, autoregulatory enhancers like ftz_up, ftz_zebra, and gt_minus1 (Hiromi et al., 1985; Hiromi and Gehring, 1987; Hoermann et al., 2016) have lower task complexity because they reinforce the expression patterns determined by other enhancers, and therefore, may not be consistent…”

“…if it consists of a cluster of TF binding sites within a short distance (Wasserman and Fickett, 1998; Berman et al., 2002; Frith et al., 2002; Halfon et al., 2002; Markstein et al., 2002; Rebeiz et al., 2002; Li et al., 2007; Wunderlich and Mirny, 2009; Hardison and Taylor, 2012).”

“To measure information in a motif, we calculated the Kullback-Leibler distance (Schneider et al., 1986; Stormo and Fields, 1998) between the motif and the composition of the intergenic regions of the genome I = L ∑ i = 1 ∑ b∈{A,C,G,T} pi ( b ) log2 pi ( b ) q ( b ) where L is the length of the…”

“…are invoked to explain why many developmental enhancers are robust to rearrangements of TF binding sites within them while some immune-responsive enhancers are intolerant to even point mutations (Thanos and Maniatis, 1995; Kim and Maniatis, 1997; Munshi et al., 2001; Arnosti and Kulkarni, 2005).”

“The 60 AP and 39 DV patterning enhancers were collected by Papatsenko et al. (2009) and provided here as Supplementary Data Sheet 1, and in this dataset, we considered the binding of 33 early patterning TFs (Fowlkes et al., 2008; MacArthur et al., 2009).”

“For example, differences in evolutionary pressures and TF cooperativity are invoked to explain why many developmental enhancers are robust to rearrangements of TF binding sites within them while some immune-responsive enhancers are intolerant to even point mutations (Thanos and Maniatis, 1995; Kim and Maniatis, 1997;…”

“…if it consists of a cluster of TF binding sites within a short distance (Wasserman and Fickett, 1998; Berman et al., 2002; Frith et al., 2002; Halfon et al., 2002; Markstein et al., 2002; Rebeiz et al., 2002; Li et al., 2007; Wunderlich and Mirny, 2009; Hardison and Taylor, 2012).”

“The 60 AP and 39 DV patterning enhancers were collected by Papatsenko et al. (2009) and provided here as Supplementary Data Sheet 1, and in this dataset, we considered the binding of 33 early patterning TFs (Fowlkes et al., 2008; MacArthur et al., 2009).”

“…if it consists of a cluster of TF binding sites within a short distance (Wasserman and Fickett, 1998; Berman et al., 2002; Frith et al., 2002; Halfon et al., 2002; Markstein et al., 2002; Rebeiz et al., 2002; Li et al., 2007; Wunderlich and Mirny, 2009; Hardison and Taylor, 2012).”

“As no TF binding sites were identified in the set of AP enhancers when using the giant PWM selected using the previous criteria, a switch was made to another available giant PWM in FlyFactor Survey with which binding sites could be predicted.”

“For example, differences in evolutionary pressures and TF cooperativity are invoked to explain why many developmental enhancers are robust to rearrangements of TF binding sites within them while some immune-responsive enhancers are intolerant to even point mutations (Thanos and Maniatis, 1995; Kim and Maniatis, 1997;…”

“We use DNase I hypersensitivity profiles to estimate the accessible regions (Thomas et al., 2011).”

“Though the well-characterized Drosophila axis patterning systems are ideal for studying how an enhancer’s regulatory tasks shape its design (Markstein et al., 2002; Papatsenko et al., 2002; Lifanov et al., 2003; Papatsenko and Levine, 2005; Halfon et al., 2008), the systems still have limitations.”

“Here, we analyzed the binding of TFs that were concurrently expressed with active enhancers based on the Berkeley Drosophila Genome Project in situ annotations (Tomancak et al., 2002, 2007; Hammonds et al., 2013).”

“…if it consists of a cluster of TF binding sites within a short distance (Wasserman and Fickett, 1998; Berman et al., 2002; Frith et al., 2002; Halfon et al., 2002; Markstein et al., 2002; Rebeiz et al., 2002; Li et al., 2007; Wunderlich and Mirny, 2009; Hardison and Taylor, 2012).”

“Enhancers can be ∼10–1,000 bps long, with a couple to tens of TF binding sites (Blackwood and Kadonaga, 1998; Yáñez-Cuna et al., 2013).”

“…if it consists of a cluster of TF binding sites within a short distance (Wasserman and Fickett, 1998; Berman et al., 2002; Frith et al., 2002; Halfon et al., 2002; Markstein et al., 2002; Rebeiz et al., 2002; Li et al., 2007; Wunderlich and Mirny, 2009; Hardison and…”

“Some PWMs were generated from DNase I footprints curated in the FlyReg database (Halfon et al., 2008); aligned versions of these footprints were not directly provided by FlyFactor Survey.”

“Though the well-characterized Drosophila axis patterning systems are ideal for studying how an enhancer’s regulatory tasks shape its design (Markstein et al., 2002; Papatsenko et al., 2002; Lifanov et al., 2003; Papatsenko and Levine, 2005; Halfon et al., 2008), the systems still have limitations.”

“Though the well-characterized Drosophila axis patterning systems are ideal for studying how an enhancer’s regulatory tasks shape its design (Markstein et al., 2002; Papatsenko et al., 2002; Lifanov et al., 2003; Papatsenko and Levine, 2005; Halfon et al., 2008), the systems still have limitations.”

“…if it consists of a cluster of TF binding sites within a short distance (Wasserman and Fickett, 1998; Berman et al., 2002; Frith et al., 2002; Halfon et al., 2002; Markstein et al., 2002; Rebeiz et al., 2002; Li et al., 2007; Wunderlich and Mirny, 2009; Hardison and Taylor, 2012).”

“Whether we use a conservative estimate of accessible regions during development (4.1Mb), a more relaxed estimate (19.4Mb), or the entire genome (175.5 Mb) (Thomas et al., 2011; Ellis et al., 2014), the same trends are seen (Figure 1C).”

“The AP axis is more complex than the DV axis, with the AP axis consisting of 14 parasegments (Nasiadka et al., 2002) and the DV axis consisting of six germ layers and sublayers (Levine and Davidson, 2005), and therefore the patterning of the AP axis requires enhancers that drive more unique gene expression patterns.”

“The 60 AP and 39 DV patterning enhancers were collected by Papatsenko et al. (2009) and provided here as Supplementary Data Sheet 1, and in this dataset, we considered the binding of 33 early patterning TFs (Fowlkes et al., 2008; MacArthur et al., 2009).”

“Therefore, we predict TF binding sites using experimentally measured binding motifs (Stormo, 2015).”

“For example, autoregulatory enhancers like ftz_up, ftz_zebra, and gt_minus1 (Hiromi et al., 1985; Hiromi and Gehring, 1987; Hoermann et al., 2016) have lower task complexity because they reinforce the expression patterns determined by other enhancers, and therefore, may not be consistent…”

“Approaches range from looking at the same enhancer across multiple species, to looking at many enhancers with similar expression or comparing genomic regions with similar binding patterns within an organism or across organisms (Erives and Levine, 2004; Ilsley et al., 2013; Li and Wunderlich, 2017; Lusk and Eisen, 2010…”

“Nevertheless, their size is within the expected size of enhancers [70, 71].”

“VISTA enhancers are mainly involved in development-related functions and are generally longer than other types of enhancers (105).”

“These overlaps ranged from 10 to 1,474 bp in length (median = 182 bp), consistent with the typical length of enhancers [10 to 1,000 bp (31)], and covered ∼3.”

“In general, eukaryotic gene expression is regulated by promoters located near the site of transcription initiation and by enhancers that are positioned on average 120 kb away from the transcription start site (TSS) of the target gene (18).”

“Previous studies have shown that enhancers can range from 50 to 1500bp long (Dao et al., 2017; Li and Wunderlich, 2017).”

“, [89,90–92]), of 3D‐chromatin conformation [93,94], and of ncRNA and eRNA expression [95,96].”

“This is not surprising since most VISTA enhancers are involved in complex embryonic development and tend to be longer than other types of enhancers [101].”

“Specifically, we calculate the Kullback–Leibler divergence of the motif from genomic background nucleotide frequencies for A/T (0.3) and GC (0.2), similar to the previously described procedure (Li and Wunderlich 2017).”

“In fact, if the target SNPs are indeed components of regulatory units, the surrounding loci are likely to be part of the units as well because the lengths of regulatory loci can be over 1000 base pairs (Blackwood & Kadonaga, 1998; Li & Wunderlich, 2017; Yáñez-Cuna et al., 2013).”

“PWMs for specific stages of fly development were retrieved from (Li and Wunderlich, 2017).”

“While current methodologies reliably probe pairwise and multi-contact DNA-DNA interactions, deciphering complex 3D chromatin organization in single cells remains challenging, particularly in the kilobase range 12-14 .”

“Finally, models trained on longer sequences (350bp) outperformed those trained on shorter ones (201bp), suggesting longer enhancers capture greater regulatory complexity and sequence context (Li and Wunderlich 2017).”

“Enhancers regulating developmental genes are longer (Li and Wunderlich 2017) and located in large regulatory domains with multiple elements controlling different functions (Pachano et al .”

“Although the 1842 bp enhancer exhibited clear transcriptional activity, its length is relatively long compared with typical functional enhancers (Blackwood and Kadonaga 1998; Li and Wunderlich 2017), prompting us to investigate whether a shorter fragment could retain enhancer activity.”