FITOM user manual
Contents
1. downstream interval and method 1 is used Default advanced options As with normal options if an option file is not provided FrToM will set options to the following default values Integrative factor true 0 F rToM will not run in integrative mode UpLeftDist 200 Farthest distance from site to compute upstream mean UpRightDist 50 Closest distance from site to compute upstream mean DwLeftDist 250 Closest distance from site to compute downstream mean DwRightDist 850 Farthest distance from site to compute downstream mean Relaxation constant 1 5 Cut off will be relaxed multiplied by 1 5 Rescoring method 2 Scores will be reevaluated multiplying by the upstream local mean ratio References Berg O G and P H von Hippel 1987 Selection of DNA binding sites by regulatory proteins Statistical mechanical theory and application to operators and promoters J Mol Biol 193 4 723 743 Berg O G R B Winter et al 1981 Diffusion driven mechanisms of protein translocation on nucleic acids 1 Models and theory Biochemistry 20 24 6929 48 Halford S E and J F Marko 2004 How do site specific DNA binding proteins find their targets Nucleic Acids Res 32 10 3040 3052 Hertz G Z G W Hartzell 3rd et al 1990 Identification of consensus patterns in unaligned DNA sequences known to be functionally related Comput Appl Biosci 6 2 81 92 O Neill M C 1989 Consensus met2. site to gene start and basic gene information For more than one sequence the results file will be divided into consecutive sequences separated by a sequence name row At the end of the file the total number of sequences and sites as well as the a priori entropy selected method and threshold are displayed Advanced options 1 Log option The log option 1 yes O no simply determines whether the program should log partial results during operation If the log option is set FrToM will generate the following files Sequence_file_name _seqs_log txt contains the read DNA sequences can be used to extract the DNA sequence in FASTA format from a GenBank source a Sequence_file_name _site_collection_log txt contains the read site collection in FASTA format a Sequence_file_name _genes_log csv contains the list of genes read from an annotated GenBank file Sequence_file_name freq_table csv contains the frequency table constructed for the read site collection and the computed positional information content Sequence_file_name _motif_score_log csv contains the list of read sites and the score associated to each of them as well as the cutoff value Sequence_file_name log displays incidences on the program run FrToM will also generate a Fitom 1log file if any problem is encountered while reading files and the program must stop 2 Method option The method option
3. the proposed hypothesis that RNA polymerase and other DNA binding proteins may use weak binding sites upstream of true promoters to improve its promoter seek dynamics as 3D diffusion appears to be too limiting to account for the efficiency of RNA polymerase in transcription Berg Winter et al 1981 Ricchetti Metzger et al 1988 Halford and Marko 2004 Thus to validate this hypothesis F1ToM was updated to compute the mean score ranking of sites in intervals both upstream and downstream of the site under evaluation so that the current site score could be corrected according to a local complexity measure that took into account the presence of nearby pseudo sites integrative correction Main operation In version 1 2 FrToM proceeds normally and scans the sequence in both strands but it does so using a look ahead method This means that FrToM will pre compute the mean score of upstream and downstream regions for the first sequence position assuming circular DNA This is called the pre run Once this initial means have been computed FrToM scans the sequence and adds the new score to the current means The site under evaluation then changes to the middle site in the mean computing interval A relaxed threshold is used to allow a substantially larger number of candidate sites which are saved together with the mean values of their surroundings Once the sequence has been scanned selected sites are re evaluated by multiplying their score wi
4. F1ToM v1 0 USER MANUAL Introduction FrToM is a computer program for the detection of binding sites in DNA or RNA sequences It implements several methods described in the literature to compute an approximation of binding affinity for a particular site based on a collection of binding sequences provided by the user Using this method FrToM scans a sequence file looking for putative binding sites across the DNA RNA sequence in both strands and filters the results according to a user specified threshold If sequence annotation is provided in the sequence file FrToM will also link the identified sites with annotated genes and it will infer their role from their location in the vicinity of genes Getting started In its current version v1 0 FrToM is a command line argument program meaning that it must be run from a DOS prompt Since it requires a file to scan and a collection of binding sites these are the only mandatory command line arguments of FrToM Thus the program can be run from the DOS prompt by typing C gt FITOM Sequence_file ext Collection_file ext The sequence file The sequence file Sequence_file ext is the file containing the sequence or sequences the user wants to scan These files can be in two possible formats FASTA and GenBank carrying the respective extensions FASTA files should contain a first definition line beginning with a gt character and the sequence thereafter starting on the s
5. allows the user to choose between the different methods described above This is a numerical parameter that in version 1 0 accepts values from 1 to 8 The different values chose the following methods Information content of a individual DNA sequence R Berg amp von Hippel Heterology Index HI Rsequence averaged Berg amp von Hippel Heterology Index Rgequence BVH Rsequence averag d differential Rsequence Rsequence Rgequence averaged individual sequence information content Regequence Ri PredictRegulon Index The index described in Yellaboina Seshadri et al 2004 for the PredictRegulon server 7 RE averaged individual sequence information content RE R The same as Rsequence Ri but using relative entropy RE instead of information content Rsequence as the averaging factor 8 Differential RE RE The same as Rsequence but using relative entropy RE instead of information content Rsequence EOR Derives from RE in the same way as Ri derives from Rsequence Described in Hertz Hartzell et al 1990 10 FitomHI RE A modification of Berg amp von Hippel heterology index HJ to truly account for the divergence between consensus and observed bases Biy Differential Regeguence Non weighted Reseguence A modification of the differential Rsequence method to converted into a non weighted method Au FP WNE 3 Background entropy option In computing the backgrou
6. d with gene information An important parameter in this tagging in order to further filter the results provided by F1ToM is to determine whether a particular site may or may not be an operator i e a site involved in promoter regulation Since known prokaryote operator sequences fall within a range of the translational start point FrToM relies on two user provided parameters the distance to promoter and the distance in gene presets The first makes reference to the maximum distance a site can be upstream of the gene translational start point in order to tag it as operator The default is 350 bp 6 Distance in gene option This second distance parameter see above makes reference to the maximum distance a site can be downstream of the gene translational start point in order to tag it as operator The default in this case is 100 bp 7 Sorting method option Results can be sorted according to two main parameters of identified binding sites their position within the genome 0 or their ranking 1 Results can later be resorted in Excel 8 Remove redundant matches option Since the sequence is scanned in both strands for palindromic motifs a same site will be located in both scans In some cases if the site is particularly good both versions will be picked up as positive results leading to duplicity of results The remove redundant matches option allows the user to remove 1 or not 0 these duplicate sites Duplicate removal p
7. e reduction in uncertainty we experience once we know that a protein or other element binds to a sequence Prior to binding our uncertainty about what bases occupy the different positions of a sequence is maximal and dictated by the base composition of the genome Once we know that the protein associated with the provided motif binds that sequence however we have much less uncertainty about what bases occupy the different positions We still have uncertainty because protein binding is a noisy issue but we have decreased our uncertainty and thus we can say we have gained information Conversely seen from the point of view of a genome the information content can also be seen as the loss of entropy at certain regions in the genome from an initial random state to a state of fixation of conserved binding sites Thus motif information content can also be as an index of the level of redundancy RJ in the different positions of the motif O Neill 1998 Even though without a complete theoretical justification a different index termed relative entropy RE has been proposed to substitute the RI in cases of heavily skewed genomes P S RE 1 AG on aS Relative entropy Schneider Stormo et al 1986 is also computed by F1rToM and can be used in different ranking methods Ranking methods Rsequence tells us how much information our motif conveys but it does not provide answers to how well a particular sequence fits in the mo
8. e relaxed cutoff are re evaluated using a correction factor that takes some of the computed means into account Three different modes of score correction are possible upstream mean global mean upstream mean local mean upstream mean downstream mean In all of them the original score is multiplied by one of the above fractions UpLeftDist DwRightDist Pe UpRightDist gt DwLeftDist sel 11 Further distance of current site to compute upstream mean X gt X Y This parameter UpLeftDist controls the farthest distance from site under evaluation in which the user wants to compute the upstream mean 12 Closer distance of current site to compute upstream mean Y gt X Y This parameter UpRightDist controls the closest distance from site under evaluation in which the user wants to compute the upstream mean 13 Further distance of current site to compute downstream mean I gt LJ This parameter DwRightDist controls the farthest distance from site under evaluation in which the user wants to compute the downstream mean 14 Closer distance of current site to compute downstream mean J gt LJ This parameter DwLeftDist controls the closest distance from site under evaluation in which the user wants to compute the downstream mean 15 Relaxation constant This parameter controls the degree to which the normal threshold specified either as relative or lite
9. econd line and typically wrapped on 70 char lines FrToM does not require 70 char wrapping but does require that the first line be strictly the definition line If the FASTA file comes from NCBI F ToM will try to extract locus and name information from the definition line If it does not succeed it will use the whole definition line as both locus and name for the organism sequence under study GenBank files contain both the sequence and annotation information and they should be used preferentially as their analysis will convey additional information on the putative function of identified sites FrToM will infer file type from the file name extension Accepted extensions are FAS FNA for FASTA files and GBK GB for GenBank files If your file does not carry any of these extensions please rename it accordingly before use Only once sequence per file is accepted for GenBank format while multiple sequences per file are allowed in FASTA format The collection file The collection file Collection_file ext is the file containing the collection of known binding sites that the user provides the program with in order to construct its model of binding site or motif Collection files can be either bare site files with aligned sites on consecutive lines or FASTA files in which each site line is preceded by an identification line beginning with gt As in the case of sequence files FrToM will infer file type from the file name ext
10. ension Accepted extensions are FAS FNA for FASTA files and TXT for bare site files If your file does not carry any of these extensions please rename it accordingly before use Specifying options Running FrToM with only two command line arguments as described above will issue a user warning stating that no options file name was provided and that the program will be running in default options mode An options file can be made available to FrToM by setting it as a third command line argument C gt FITOM Sequence_file ext Collection_file ext Options_file ext The options file All options can be set in the options file The options file options_file ext is a simple text file containing all the advanced options If a sample options file was not present with your distribution please contact the developers Options are set simply by assigning each parameter with a positive integer value in a single line preceded by a sign All other text is considered to be comments and is discarded by the program A sample of the options file can look like this OPTIONS 1 Log results 1 yes 0 no 1 2 Method 1 8 see documentation for details 5 where the first option boolean log would be set to 1 yes indicating that file logging is to be conducted while the second option method is set to 5 All options can be set this way following always the same ordering which is the one used in describing each option in
11. hods for finding and ranking DNA binding sites Application to Escherichia coli promoters J Mol Biol 207 2 301 10 O Neill M C 1998 A general procedure for locating and analyzing protein binding sequence motifs in nucleic acids Proc Natl Acad Sci U S A 95 18 10710 5 O Neill M C 2003 A procedure for identifying loosely conserved protein binding DNA sequences Methods Enzymol 370 229 37 Ricchetti M W Metzger et al 1988 One Dimensional Diffusion of Escherichia coli DNA Dependent RNA Polymerase A Mechanism to Facilitate Promoter Location Proc Natl Acad Sci U S A 85 13 4610 4614 Schneider T D 1997 Information Content of Individual Genetic Sequences Journal of Theoretical Biology 189 4 427 441 Schneider T D G D Stormo et al 1986 Information content of binding sites on nucleotide sequences J Mol Biol 188 3 415 31 Yellaboina S J Seshadri et al 2004 PredictRegulon a web server for the prediction of the regulatory protein binding sites and operons in prokaryote genomes Nucleic Acids Res 32 Web Server issue W318 20
12. ive sites and save only what it considers true positives This is done using one of the ranking methods described above and a threshold that again is provided by the user Since the specific cut off value depends on the method it is often more convenient to specify it using a relative threshold This value in the 0 1 range specifies how many sites from the original collection should be above the cut off value Therefore if the relative threshold is 0 5 a cut off value will be chosen by FrToM so that assuming a normal distribution only 50 of the collection sites is above the threshold If available in the sequence GenBank file FrToM will also gather information on the location of genes in the genome Using this information and the position of identified binding sites F1ToM will correlate the results in order to link sites to genes Following user defined limits FrToM will assign different tags to sites labeling them as operators within limits intragenic within coding region or intergenic between genes Results FrToM will save the results of the analysis into a comma separated value CSV file Sequence_file_name _ method _ threshold csv that can be opened directly with spreadsheet software such as Ms Excel The results file contains the identified binding sites together with their score position and strand In case gene information is available the results file will also contain a site tag the distance from
13. nd entropy Hbefore several authors have proposed assuming equiprobability Hbefore 2 bits irrespective of the genome composition on the argument that a protein does not know about genome composition in skewed genomes and hence its a priori uncertainty should be assumed to be maximal Schneider Stormo et al 1986 Even though the argument can be disputed mainly because the protein will have also evolved in the skewed genome F1rToM allows the user to specify whether to use a fixed 2 bits background entropy 0 or to derive it from genome composition 1 In case of a FASTA file with multiple sequences FrToM will assume that Hbefore is 2 bits regardless of this option s setting 4 Threshold The value introduced in the threshold option can specify two different parameters depending on the setting of option 9 If option 9 is set to O no literal cut off FrToM will use the value specified in this option 0 1 as a threshold relative to the collection of sites provided by the user FrToM will assume that the provided collection of binding sites has a normal distribution of scores and will determine the method dependent cut off that selects the percentage of sites from the collection indicated by the user through this parameter If option 9 is set to 1 literal cut off FrToM will use the value provided in option 4 as the cut off for the chosen method 5 Distance to promoter option If available identified sites are tagge
14. on about the redundancy of each position the information from the rest of bases at that position discarded by both methods is not used To correct this O Neill proposed averaging this kind of methods with the know redundancy index of the collection O Neill 1989 so that the final score was given by L l BvH 1 R egene R gt R oicice D R D 1 L R sequence BYH X Rosa l 1 An even more interesting approach was also proposed by O Neill to solve some of the different problems of the methods described above The differential Rsequence R sequence iS a method to score putative binding sites based on the successive evaluation of the Rsequence before Rsequence and after Rsequence assuming that the site is a member of the collection R egnens D R sniene J RN J R guene l Computing then the difference between both Rsequence values provides a measure of how well does the putative site sit within the collection If the site does not agree with the motif profile generated by the collection the expanded Rsequence Rsequence will decrease with respect to Rsequence and the difference will be negative Conversely if the site agrees well with the motif profile Rsequence will increase and the difference will be positive Cut off and tagging A priori every subsequence of the same size as the motif is a putative binding site Therefore as Fr1ToM scans the sequence it must discard false posit
15. ral cutoff is relaxed in integrative mode Threshold relaxation is required for the integrative mode to introduce new information to ranking Else the method would only re rank the sites already detected in normal operation The default relaxation constant is 1 5 Bigger relaxation constants will provide the integrative mode with additional freedom to choose sites based on their integrative component a larger pool of candidate sites will be rescored while smaller constants will reduce the contribution of the integrative factor Care should be taken in specifying very big gt 2 relaxation constants when analyzing large files since this could result in a very large amount of sites selected for rescoring which may lead to slow runs and in extreme cases out of memory errors 16 Rescoring method The rescoring method option 1 3 allows the user to specify the method that will be applied to rescore candidate sites in integrative mode Three methods are available in FrToM to rescore sites which differ on the fraction they apply as a multiplicative correction factor 1 Upstream mean global mean 2 Upstream mean local mean 3 Upstream mean downstream mean By combining this option with interval definition options 11 14 many different re evaluation strategies may be assayed For instance if one is interested in obtaining the local global mean a suitable upstream region is defined as the intended local mean regardless of the
16. roceeds based on a best ranking policy among the two versions of the same site only the better scoring one will be preserved 9 Literal cut off option This option allows the knowledgeable user to introduce a literal cut off instead of a relative percentual threshold If the option is set 1 will use FrToM the value set in the threshold option as a literal cut off instead of as a relative threshold Default options If an options file name is not specified FrToM will use a set of predefined option values These default values are listed below Log option true 1 Additional log files will be saved Method RI averaged BvH 3 Default method Background entropy false 0 Background entropy not computed Threshold 0 8 relative threshold Threshold at 80 of collection sites Promoter distance 350 bp 350 bp max distance upstream of ATG for operator site In gene distance 100 bp 100 bp max distance downstream of ATG for operator Sort option false 0 Sort results by genomic position Remove redundant matches true 1 Redundant matches will be removed Literal cut off false 0 Threshold value will be used as relative threshold CHANGES IN VERSION 1 2 Version 1 2 is a major update of FrToM to take into account local complexity in the form of signal overrepresentation in the detection of binding sites The idea developed initially for bacterial promoter detection is based on
17. th a correction factor derived from a ratio between means e g upstream downstream mean Additional advanced options The integrative factor correction as it is called in FrToM described above can naturally be turned on or off and is completely user customizable The following describes the different options added to FrToM and their role in program operation 10 Integrative factor This option indicates whether FrToM operates in normal 0 or integrative 1 mode When in normal mode FrToM works in the way explained above for version 1 0 In integrative mode FrToM computes four mean values local mean max X Y max X Y upstream mean X x downstream mean y Y and the global mean for all the sequence The intervals can be freely specified by the user For instance one may compute the upstream mean from positions 200 to 50 of the site and the downstream mean from positions 100 to 500 from site The only restriction is that the intervals may not overlap The local mean interval is always site symmetrical and covers the maximum distance from the site specified by the user in the above example it would cover the 500 500 interval After computing the pre run mean F1ToM starts analyzing the genome with a look ahead of max X Y In this mode of operation the cut off value usually applied to putative sites is relaxed using a user determined constant After scanning the sequence the sites with scores above th
18. the following pages Main operation Before going into the description of the advanced options it is interesting to describe the main modus operandi of the program As mentioned above F1ToM starts by loading the provided sequence and collection files Position specific weight matrix and information content The collection file is then used to compute the motif position specific weight matrix PSWM This matrix is a matrix of the relative frequencies of each nucleotide at each position in the motif 1 2 3 4 5 6 A 0 031 0 055 0 650 0 349 0 309 0 007 c 0 928 0 015 0 015 0 071 0 158 0 007 G 0 007 0 206 0 166 0 031 0 079 0 976 T 0 031 0 722 0 166 0 547 0 452 0 007 Consensus C T A T T G From the positions specific weight matrix the information content Rsequence or redundancy index of the motif can be computed according to the following formula Rene LD X H pore H pa D EU Srog A9 gt p S log p S l 1 l 1 SEQ SEQ StS frequency of base S in the genome p S frequency of base S in the motif PSWM F ToM adds 101 to each motif frequency to avoid O log 0 terms for bases not represented in the collection before a priori entropy Harter entropy after binding as described by Schneider et al Schneider Stormo et al 1986 and based on the assumption of positional independency The information content of a motif tells us about th
19. tif profile which is what is required to scan for and rank putative binding sites Several ranking methods have been proposed with diverse degrees of theoretical justification FrToM provides two basic scoring methods that can be used to rank putative binding sites The sequence information content Ri Schneider 1997 is a method derived from the information content Rsequence formula that scores each position of a particular site j based on ratio of frequency in the motif with respect to genomic frequency for the particular base observed in the site Sea 1 4 N R 1 D P S S dog ro tog 2824 1 N 1 4 N is the zero frequency correction following Laplace Law of Succession to estimate the frequency of a base present in a sequence and not present in the collection N is the number of sequences in the collection Another proposed method is the Berg von Hippel Heterology Index HJ based on the relative frequency of the observed base with respect to the dominant consensus base at each position Berg and von Hippel 1987 L BvH HI In Pfam HUN mi Pll 1 N obs 1 N is again a correction factor for bases with zero frequencies in the collection As a result of their formulation both ranking methods discard information from the other motif base frequencies As explained in O Neill 2003 this can lead to erroneous scoring where the same score may be given to little or heavily conserved positions since informati
Download Pdf Manuals
Related Search