ikmb/esga repository overview

⁠ESGA - Genome Annotation (v2)

or

This pipeline peforms annotation of novel genomes using a combination of evidence alignments, evidence-based gene building and ab-initio gene building.

It does not constitute a novel algorithm but combines many well-established tools into an easy-to-use "one click" solution that is accessible even to inexperienced users.

⁠Pipeline main steps

The minimum requirements are a genome file and at least one type of evidence.

From this, the pipeline can run the following processing steps:

• Repeat-mask the assembly
• align species-specific proteins to build protein-based gene models
• align proteins against a genome and generate annotation hints
• align transcripts against a genome and generate annotation hints
• align RNA-seq reads against a genome and generate annotation hints
• assemble transcripts from aligned RNA-seq reads and generate annotation hints (Trinity pipeline)
• Produce evidence-based gene models from aligned transcript sequences (PASA pipeline)
• Produce ab-initio, hint-supported gene models (AUGUSTUS pipeline)
• Produce consensus annotation from all of the above (EvidenceModeler pipeline)
• Predict non-coding RNA loci using RFam

Optional:

• Train a novel ab-initio prediction profile for AUGUSTUS (using PASA transcripts or SPALN protein models)

⁠Test data

A simple test data set can be downloaded here⁠

⁠Documentation

Documentation about the pipeline can be found in the docs/ directory or under the links below:

1. What happens in this pipeline?⁠
2. Recommendations⁠
3. Installation and configuration⁠
4. Running the pipeline⁠
5. Output⁠
6. Troubleshooting⁠
7. What's next⁠

⁠Why ESGA?

ESGAs' implementation was motivated by our need for a "fully contained" annotation solution that uses current programming paradigms and structures to rapidly annotate a metazoan genome. For us, this means: modular, open code as well as support for a range of popular schedulers to translate the workflow to a compute farm. In addition, we wanted to support input data that is missing from some other available solutions - including raw RNAseq reads and related genomes/annotations to lift-over annotations from other model systems. Finally, we wanted all of this to be usable without having to go through lengthy (and often rather painful) installation procedures for a host of software tools. ESGA uses a small number of docker containers to deliver all the required dependencies. In our experience, this solves some of the main issues people typically experience when trying to annotate a genome.

All you need to do is write a small config file and provide singularity/docker support on your cluster. Done.

⁠Credits

This pipeline was written by Dr. Montserrat Torres (MontseTor⁠) and Dr. Marc Höppner (marchoeppner⁠) at IKMB⁠. The authors gratefully acknowledge inspiration, fruitful discussions and a few useful code snippets from the nf-core⁠.