Home
TransATor is based on the correlation between the phylogenetic origin of the keto-acylsynthase (KS) domain and its substrate specificty. We recommend to upload the entire amino acid information of a given gene cluster which will result in a complete annotation of core biosynthetic enzymes. Individual KS domains spanning at least the region between the EDAGY and HGTGT motif are, however, sufficient for reliable substrate predictions.
Figure 1: Phylogenetic tree showing KS sequences clades.
All 657 trans-AT PKS KS sequences of all annotated trans-AT PKS gene clusters were retrieved from public databases and aligned using Muscle. The alignment was refined using MAFT. A RAxML phylogentic tree was constructed and manually grouped into 48 clades according to their substrate specificty (see Fig. 1). HMM models were generated for each individual clade. In addition, alignments were constructed from individual enzymatic domains commonly present in trans-AT PKS and NRPS biosynthetic gene clusters. Moreover, FUZZPRO patterns are used to distinguish elongating from non-elongating (KS0) KSs, to determine the stereospecificity of KR domains and to distinguish PS from DH domains. The NRPSpredictor2 was incorporated to verify the specificity of amino acids incorporated by NRPS modules.
In a first step, a given sequence is annotated based on the comparison with the indivual HMM models for KS substrate specificties and all other present core enzymatic domains (see Fig. 2). Based on these annotations, the polyketide core structure is assembled from the determined KS substrate specificity. KS-monomers, representing the alpha-gamma position of a biosynthetic intermediate are connected and the structure is verified by the presence or absence of essential enzymatic domains for a certain monomer to be formed (e.g. a branching domain for glutarimide-containing polyketides or a PS for pyran ring-containing polyketides). The structure is then refined by the additon of amino acid side chains, cyclization patterns or the incorporation of stereochemical information. In cases in which the enzymatic machinery does not terminate with an NRPS module or a non-elongating KS, the reductive state of the beta-carbon of the last incorporated building block is determined by the so-called co-linearity rule used to predict the strcutures of textbook cis-AT PKS derived polyketides. (see Fig. 3).
Upload your data in proper FASTA format (Fig. 4). Fig. 5 shows the output of the bacillaene trans-AT PKS biosynthetic machinery as an example. On top the predicted polyketide core structure is displayed. For the predicted structure only the top HMM hit of the respecitive KS sequence is taken into consideration. Below, the gene cluster annotation is displayed. The uploaded sequence is displayed as a line with annotations based on HMM models (boxes) being displayed below the line and FUZZPRO patterns (diamonds) on top of the line representing the sequence. Each annotation contains information on the enzymatic domain, its substrate specificity, e-value and localisation within the input sequence. Hovering across the annotation will display this information. For KS sequences, the top 5 hits are displayed. While the structure is generated from the top hit only, it might be worth while comparing the e-values of the other hits and, if necessary, also take other monomers for the corresponding position in the predicted molecule into consideration. Output files can be exported as images.
TransATor is a genome mining tool designed for the structure prediction of trans-AT PKS derived polyketides. TransATor can be used for multiple scientific questions. First, transATor faciliates in-silico dereplication studies both on gene cluster and more importantly molecular level. TransATor can be used for the priorization for novel, unusual polyketide scaffolds and the in-silico linkage of orphan gene clusters to known polyketides. Most importantly, transATor was designed to guide the isolation and structure elucidation process of novel trans-AT PKS derived polyketides. Moreover, transATor can be used to generate and refine biosynthetic models,
Yes. A stand-alone version of trans-ATor is available for download and runs on Linux and Mac OS X operating systems.
In case your sequence was not uploaded in proper FASTA format, the output page will remain blank. You can convert your data into proper FASTA format using e.g. MultiFasta builder (batch conversion), ReadSeq at EBI or ReadSeq at NIH.
transATor is developed by Pablo Moreno and Eric Helfrich, within the groups of Christoph Steinbeck at EBML-EBI and Jörn Piel at ETH Zurich, respectively.