Database implementation

The schematic structure of the ECRbase data analysis is presented in Figure S1. The database first processes whole genome pairwise alignments of multiple vertebrate genomes available from the ECR Browser to identify Evolutionary Conserved Regions (ECRs). Currently there are over 26M ECRs available in the ECRbase that correspond to regions extracted from pairwise comparisons of all the available species. Next, these ECRs are used to determine synteny blocks that interconnect these genomes. Since the identified synteny blocks are based on nucleotide alignments, not on protein similarity, are thus capable of more accurately demarcate synteny breakpoints in long intergenic regions. This strategy potentially provides user with more accurate synteny maps with longer syntenic stretches for closely related vertebrates (such as human and mouse, for example) than those restricted to gene comparisons. In parallel to the ECR identification we've implemented the extraction of vertebrate promoters using RefSeq, knownGene, and other species RefSeq gene annotations available from the UCSC Genome browser database (Hinrichs et al. 2006,Pruitt et al. 2005). At the final step, DNA sequences for the identified ECRs and promoters undergo annotation of TFBS. Subsequently, all the processed data is collected, binned according to the corresponding genome, and distributed through the central ECRbase interface available at http://ecrbase.dcode.org. Large ECRs and TFBS files are compressed (using the gzip utility) to facilitate data download. Despite the compression, some of the files continue to be relatively large and, therefore, some users may find it helpful to use automated file download utilities for fetching data from the ECRbase. Below we summarize the details of methods employed for data extraction and generation.

Fig. 1.  Schematic pipeline for the ECRbase data analysis.

Evolutionary Conserved Regions.

ECRs are computed as regions greater than 100bps in length and greater than 70% nucleotide sequence identity (Table S1). For a region to be classified as an ECR, it is required to be present in both species. There are cases when a conserved region in one species has accumulated significant insertions in the second species and therefore, its second species conservation falls below the required threshold. Elements that exhibit this conservation pattern are excluded from the database. Stricter thresholds, of a minimum length of 350bps and conservation level of 77% ID are used for identifying conserved elements termed coreECRs – regions that are implied to have a higher probability of being functional than regular ECRs (Ovcharenko et al. 2004,Prabhakar et al. 2006). ECRbase reports genome positional information of ECRs (and coreECRs), their length and percent identity as well as the corresponding parameters for their orthologs in other genomes.

Table S1.  Number of ECRs in inter-species alignments for the human (hg18), dog (canFam2), mouse (mm8), rat (rn4), chicken (galGal3 & gg2), frog (xt4), and Fugu (fu4) genomes (in thousands)

	Dog	Mouse	Rat	Chicken	Frog	Fugu
Human	2,521	1,289	1,189	200	120	73
Dog		1,042	972	178	115	71
Mouse			2,311	169	109	74
Rat				162	107	70
Chicken					117	67
Frog						73

Synteny.

Synteny between vertebrate genomes is determined as previously described (Ovcharenko et al. 2005). Briefly, we use sets of 3 consecutive ECRs (two neighboring ECRs are selected as ‘consecutive’ if they are separated by <100kb in both genomes) to define anchors of inter-genome synteny. These synteny anchors are used to construct larger synteny blocks by clustering ECR triplets from matching chromosomes using the same maximum 100kb separation threshold (Table S2). Since a great number of genomes are available in draft sequence format (in a multi-scaffold configuration), several artificial synteny breakpoints originate simply from the scaffold edges prematurely disrupting the syntenic structure. Short scaffolds can also potentially prevent the identification of the 3-ECR synteny anchors thus also leading to the elimination of some synteny relationships and/or generation of incomplete syntenic blocks. For that reason, synteny assignments originating from unfinished genomes should be treated with caution.

Promoters.

ECRbase utilizes RefSeq, knownGene, and ‘other species RefSeq’ gene annotation available at the UCSC Genome browser database (Hinrichs et al. 2006,Pruitt et al. 2005) to localize the genomic position and the strand of gene transcripts in vertebrate genomes. Overlapping transcripts are combined into unique genes and the outermost 5’ end is used as a landmark of transcriptional start site (TSS). Alternative promoters that are exonic, intronic or in untranslated regions can not be determined by this analysis and are therefore not represented in the database. Next, the data extraction utility selects the 1.5kb region upstream of the gene TSS, annotates it as the promoter element and automatically fetches the corresponding DNA sequence (repetitive elements are indicated by lower-case letters consistent with data representation in the UCSC Genome browser). Promoter elements are limited to intergenic spaces and are dependent on the location of neighboring genes. In cases where the intergenic region is significantly shorter than 1.5kb, the identified promoters span the entire intergenic space between the two transcripts and are therefore less than 1.5kb. ECRbase reports positional and directional information of promoters as well as it provides the name of the gene the promoter is associated with. Bi-directional promoters (promoters shared by two genes transcribed in a head-to-head manner) are reported twice – once for each transcript.

Table S2.  Longest syntenic block size from inter-species comparison of the human (hg18), mouse (mm8), chicken (gg2), frog (xt4), and Fugu (fu4) genomes (in thousand basepairs, kb)

Second Base	Human	Mouse	Chicken	Frog	Fugu
Human	-	56,101	48,475	9,233	9,656
Mouse	54,507	-	38,134	7,888	8,783
Chicken	19,105	16,106	-	7,952	4,704
Frog	5,479	5,333	6,529	-	3,529
Fugu	1,990	1,723	1,539	1,397	-

Transcription factor binding sites.

We utilize TRANSFAC Professional database of PWM corresponding to vertebrate transcription factors (version 9.4) (Matys et al. 2006) to map candidate TFBS in genomic sequences. TFBS are mapped as previously reported, using the tfSearch (Ovcharenko et al. 2005) utility that employs a suffix tree technique to rapidly identify motifs in DNA sequences. In an effort to limit the number of false positive TFBS predictions we avoid using default PWM sequence similarity parameters, but instead perform an independent optimization of thresholds for different TFBS that warrants 5 or less TFBS predictions per 10kb of random sequence. Each ECR and promoter element undergoes a TFBS mapping, and positional and directional information of each TFBS inside these elements. TFBS data is next collected and distributed through the corresponding portal of the ECRbase.

References.

Ovcharenko, I., Nobrega, M.A., Loots, G.G. and Stubbs, L. (2004) ECR Browser: a tool for visualizing and accessing data from comparisons of multiple vertebrate genomes. Nucleic Acids Res, 32, W280-286.

Hinrichs, A.S., Karolchik, D., Baertsch, R., Barber, G.P., Bejerano, G., Clawson, H., Diekhans, M., Furey, T.S., Harte, R.A., Hsu, F. et al. (2006) The UCSC Genome Browser Database: update 2006. Nucleic Acids Res, 34, D590-598.

Pruitt, K.D., Tatusova, T. and Maglott, D.R. (2005) NCBI Reference Sequence (RefSeq): a curated non-redundant sequence database of genomes, transcripts and proteins. Nucleic Acids Res, 33, D501-504.

Ovcharenko, I., Stubbs, L. and Loots, G.G. (2004) Interpreting mammalian evolution using Fugu genome comparisons. Genomics, 84, 890-895.

Prabhakar, S., Poulin, F., Shoukry, M., Afzal, V., Rubin, E.M., Couronne, O. and Pennacchio, L.A. (2006) Close sequence comparisons are sufficient to identify human cis-regulatory elements. Genome Res, 16, 855-863.

Ovcharenko, I., Loots, G.G., Nobrega, M.A., Hardison, R.C., Miller, W. and Stubbs, L. (2005) Evolution and functional classification of vertebrate gene deserts. Genome Res, 15, 137-145.

Matys, V., Kel-Margoulis, O.V., Fricke, E., Liebich, I., Land, S., Barre-Dirrie, A., Reuter, I., Chekmenev, D., Krull, M., Hornischer, K. et al. (2006) TRANSFAC and its module TRANSCompel: transcriptional gene regulation in eukaryotes. Nucleic Acids Res, 34, D108-110.

Ovcharenko, I., Loots, G.G., Giardine, B.M., Hou, M., Ma, J., Hardison, R.C., Stubbs, L. and Miller, W. (2005) Mulan: multiple-sequence local alignment and visualization for studying function and evolution. Genome Res, 15, 184-194.