Table of Contents Use GC3Pie to annotate your large sequence dataset. Configure TRAL. Prepare your data. Sequence data Sequence profile model to sequence data mapping Sequence profile models pyfaidx Test the file preprocessing. Perform a test run. Install GC3PIE. Usage Configure TRAL. Prepare your data. Sequence data Sequence profile model to sequence data mapping Sequence profile models pyfaidx Test the file preprocessing. Sequence profile model to sequence data mapping Sequence profile models Test the file preprocessing. Perform a test run. Install GC3PIE. Perform overlap filtering of redundant tandem repeat annotations. Searching for particular repeats This tutorial will work you through an example of annotating tandem repeats on a very large sequence data set: We will annotate tandem repeats with sequence profile models, and with de novo detection algorithms, perform significance testing and overlap filtering all in one. For very large sequence sets (> 6000 sequences ~ 10h runtime), you may wish to perform the annotation on several computing nodes in parallel. Here, we provide a system to automise the distribution and collection of tandem repeat annotation jobs with GC3Pie . Everything is explained with a toy workflow located in the TRAL source . You can print the path to the local copy included in TRAL with this Python code: To follow this tutorial, set a shell variable $MYTRAL to the package variable PACKAGE_DIRECTORY. In bash and sh, this can be done with: For csh or tcsh, use: Configure TRAL. ¶ We start by adapting the TRAL configuration files to choose for example which external tandem repeat detection algorithms you would like to run on the sequence data. Don’t forget to add a comma behind the last entry in this list, otherwise the file format is violated. Now, let’s move ahead to data acquisition: Prepare your data. ¶

To annotate sequences with tandem repeats, we already acquired the following data and stored it in $MYTRAL/examples/workflow : Sequence data ¶ We downloaded all UniProtKB/Swiss-Prot human PRDM genes in fasta format… … and split the sequence file into subsets on which the tandem repeat annotation can be performed in parallel. On a sidenote, fastasplitn is a useful tool for splitting large data files. Sequence profile model to sequence data mapping ¶ We downloaded the mapping to PFAM annotations for these sequence also from www.uniprot.org. Sequence profile models ¶ We downloaded all needed PFAM sequence profile models, for example of the Zinc finger . The toy example requires the Python package pyfaidx for .fasta file indexing. Test the file preprocessing. ¶ To allow fast access to the data for the tandem repeat annotation steps, it is transformed to Python pickle objects: On the small toy example, this file preparation step should run fast. Perform a test run. ¶ With the following command, annotation is performed on uniprot_PRDM_1.fasta : If this runs fine, you should see annotation results in: Now, we can move ahead to automated distributed annotation with GC3PIE. Install GC3PIE. ¶ You can follow the official GC3Pie installation instructions . Upon installation, GC3Pie needs to be locally configured . Adapt this command to run the tandem repeat annotation workflow ( more information ): Control the workflow with GC3Utils .