1. Introduction

This program analyses a pair of homologous mammalian DNA sequences (i.e. human and mouse/rat) for high probability binding sites of known transcription factors. A set of Position-Specific Scoring Matrices (PSSM) has been carefully constructed from mammalian transcription factor binding sites deposited in TRANSFAC database. These matrices are species-specific (i.e. human and mouse/rat) and class-specific (human/mouse/rat). Matrices were created using program WConsensus (Hertz and Stormo, Bioinformatics 1999).

FOOTER is a statistical algorithm that employs phylogenetic footprinting principles to identify the most likely human-mouse pairs of binding sites. The algorithm evaluates each pair of sites using two tail probabilities:

1) Relative distance of the two sites from the Transcription Start Site (TSS) or the closest boundary of conserved region.

2) The relative PSSM score.


Tail probabilities are calculated based on the hypothesis that the compared pair of sites occur by chance and they are not true binding sites. The weighted sum of these tail probabilities constitutes the FOOTER score and a Weighted Average p-value (WAP) is calculated based on this score.

The program presents the results in table format. The results are consisted of the transcription factor name, the identified human site and its position/orientation, the identified mouse site and its position/orientation, the FOOTER combined score and the weighted average p-value. The output also provides links to the human and mouse analyzed promoter sequences and alignment output.

 

2. Sequence Input

The user has the option to specify in the input either two DNA squences (typpically promoters that can include introns, exons, etc) or one protein sequence (human or mouse/rat). In the first case, the two DNA sequences are aligned first (program DBA) and searches for high-scoring motifs and then FOOTER algorithm selects the best motifs according to its scoring scheme. If a protein sequence is defined, then the the program will try to identify the TSSs by searching the corresponding Unigene collections for the longest mRNAs. Once the TSSs have been identified, the program extracts a user-defined length aroound them and calls this "promoter". The extracted sequence might include UTR, introns, exons, etc. The user is cautioned, though, that correct identification of the TSS is not always possible using that way! When the promoter sequences are known, it is always better to use input option #1. All submitted sequences should be in FastA format.

Search length:
The user can specify the length of upstream and downstream sequences that wishes to be analyzed. Lengths are valid only if the user submits a protein sequence.

3. Parameters

a) PFd Weight:
The user can define the weight given to the probability for the relative distance of the putative sites from the TSS or the closest conserved region boundary.  For more details on the probability score, see (Corcoran et al.).

b) PFs Weight:
The user can define the weight given to the probability for the relative PSSM scores.  For more details on the probability score, see (Corcoran et al.).

c) WAP Cut-off:
The user specifies the weighted average p-value cutoff.

d) Seed:
The user can define the number of seed putative binding sites per 'n' base pairs.

4. Output

FOOTER predictions are presented in a table format, preceeded by a PNJ file showing the degree of conservation (color coded) between the two regions analysed. The columns of the table contain information about the transcription factor(s) predicted to bind, the sites' sequence and position and information about the FOOTER combined score and WAP value.


a) Factor:
The name of the transcription factor predicted to bind in this region.

b) Human site:
The transcription factor binding site sequence in the human promoter.

c) PosH (orient.):
The position where the human site found and its orientation. Note that position 1 is the beginning of the specified sequence, not the Transcription Start Site (TSS). Orientation is determined with respect to the specified sequences.

d) Mouse site:
The transcription factor binding site sequence in the mouse promoter.

e) PosM (orient.):
The position where the mouse site found and its orientation. Note that position 1 is the beginning of the specified sequence, not the Transcription Start Site (TSS). Orientation is determined with respect to the specified sequences.

f) Score:
The FOOTER combined score, which is the negative sum of the log-probabilities of the scores for the individual criteria (Corcoran et al.). Practically, this score is equal to -ln(WAP). Score values around 5 are giving the best results in most cases.

f) WAP value:
The Weighted Average p-value as it is calculated by FOOTER.

5. Reference

Please reference this work as follows:

Corcoran et al. "FOOTER: a quantitative phylogenetic footprinting method for efficient recognition of cis-regulatory elements" Genome Res (2005) in press.

 

6. Transcription Factors

The user can select which transcription factors should be considered in the search.  Currently, FOOTER offers 127 different factors.  We are currently working on adding the ability for the user to input their own count matrix.  Note: The 'check all' and factor group check do not currently work on all browsers.

 

7. Sample Protein Sequence

>Mouse_PEPCK
       mppqlhngld fsakviqgsl dslpqavrkf vegnaqlcqp eyihicdgse eeygqllahm
       qeegvirklk kydncwlalt dprdvaries ktviitqeqr dtvpipktgl sqlgrwmsee
       dfekafnarf pgcmkgrtmy vipfsmgplg splakigiel tdspyvvasm rimtrmgisv
       lealgdgefi kclhsvgcpl plkkplvnnw acnpeltlia hlpdrreiis fgsgyggnsl
       lgkkcfalri asrlakeegw laehmlilgi tnpegkkkyl aaafpsacgk tnlammnpsl
       pgwkvecvgd diawmkfdaq gnlrainpen gffgvapgts vktnpnaikt iqkntiftnv
       aetsdggvyw egideplapg vtitswknke wrpqdaepca hpnsrfctpa sqcpiidpaw
       espegvpieg iifggrrpeg vplvyealsw qhgvfvgaam rseataaaeh kgkiimhdpf
       amrpffgynf gkylahwlsm ahrpaaklpk ifhvnwfrkd kdgkflwpgf gensrvlewm
       fgriegedsa kltpigyipk enalnlkglg gvnveelfgi skefwekeve eidryledqv
       ntdlpyeier elralkqris qm

Factors to Try with the Sample Sequence:

Under Leucine Zipper Factors:

CREB and C/EBP-beta