Seq2Logo-2.0: Release notes: 22 March 2014 ================================================================================ * Seq2Logo has been completely rewritten to make it easier maintainable and allow more flexibility. * Most of the script now uses a faster algorithm which should be noticeable for those creating long motifs. * Enabled the use of PSSM inputs with 2-50 columns, compared to the limit of 2-20 amino acids previously. This was done to allow the use of a column for showing gaps in the PSSM motif primarily. * Fixed a limitation of PSSM inputs where a PSSM with no position number and where the weights were presented as integers resulted in the first column to be interpreted as the position number. Now Seq2Logo relies on the header line containing the characters to determine if the first position is a position number or a weight. * Reduced alphabet limitations. It is now possible to use more symbols and not only the amino acids symbols. Note: Remember to upload a new backgound frequency table, and also a new substitution matrix if you use pseudo counts. * PSSM inputs can now use all ASCII symbols available. * Fixed poscript scaling and repositioning for all symbols. Now all symbols should fit their appointed bounding boxes. * Fixed a Postscript zero division problem occuring when there were no positive values in the PSSM. * Fixed a Postscript zero division problem caused by small height values. * Integration of DNA sequence input interpretation and calculations. It is now possible to upload DNA sequence alignments and get a sequence motif using Seq2Logo. Note: Seq2Logo does neither have any default transition table or reference distributions for nucleotides, so the user should either provide these themselves or set beta to 0 and chose the Shannon logo. * A more flexible and intuitive color tool has been added, giving the user more power over the visual setup of the motif. * Thorough documentation has been added to the source code, the commandline interface has been updated, and these realease notes are now available from the web. * The Hobohm 1 algorithm has been upgraded to sort the alignment sequences so the sequences with most data (eg. least amount of gaps) are first, and those with many gaps are added in the end. This fixes an issue where a sequence with many gaps attracted to many sequences to its cluster, and where these sequences in the final cluster could be quite dissimilar. * The Blosum matrix and background frequency file upload can now also be copy/pasted just like the file upload, by clicking the blue 'switch' link. * An option for adding the consensus sequence to the x-axis position numbers is now available in the advanced menu. * The settings menu on the output page has had a small change. It is now possible to download the used BLOSUM substitution matrix and the background frequencies. * The use of a position specific reference distribution (background frequencies) is now possible. The reference distribution format has been adjusted so it works with the same setup as PSSM input files. Note: To specify a global distribution for all positions, just upload a file containing only one distribution line (i.e. first position). Additionally, any missing positions or positions containing a 0 or a negative value, will be overridden to use a flat/uniform distribution. A user warning is recorded and can be found in the warning section of the output page if this is true. * It is now possible to pipe the input to Seq2Logo instead of using the -f argument. See the readme for an example hereof.