BRAKER3: Fully automated genome annotation using RNA-seq and protein evidence with GeneMark-ETP, AUGUSTUS, and TSEBRA

Abstract

Gene prediction has remained an active area of bioinformatics research for a long time. Still, gene prediction in large eukaryotic genomes presents a challenge that must be addressed by new algorithms. The amount and significance of the evidence available from transcriptomes and proteomes vary across genomes, between genes, and even along a single gene. User-friendly and accurate annotation pipelines that can cope with such data heterogeneity are needed. The previously developed annotation pipelines BRAKER1 and BRAKER2 use RNA-seq or protein data, respectively, but not both. A further significant performance improvement integrating all three data types was made by the recently released GeneMark-ETP. We here present the BRAKER3 pipeline that builds on GeneMark-ETP and AUGUSTUS, and further improves accuracy using the TSEBRA combiner. BRAKER3 annotates protein-coding genes in eukaryotic genomes using both short-read RNA-seq and a large protein database, along with statistical models learned iteratively and specifically for the target genome. We benchmarked the new pipeline on genomes of 11 species under an assumed level of relatedness of the target species proteome to available proteomes. BRAKER3 outperforms BRAKER1 and BRAKER2. The average transcript-level F1-score is increased by about 20 percentage points on average, whereas the difference is most pronounced for species with large and complex genomes. BRAKER3 also outperforms other existing tools, MAKER2, Funannotate, and FINDER. The code of BRAKER3 is available on GitHub and as a ready-to-run Docker container for execution with Docker or Singularity. Overall, BRAKER3 is an accurate, easy-to-use tool for eukaryotic genome annotation.

Figure 1.

Schematic view of the BRAKER3 pipeline. Required inputs are genomic sequences, short-read RNA-seq data, and a protein database. The RNA-seq data can be provided in three different forms: IDs of libraries available at the Sequence Read Archive (Leinonen et al. 2010), unaligned reads, or aligned reads. If library IDs are given, BRAKER3 downloads the raw RNA-seq reads using the SRA Toolkit (https://trace.ncbi.nlm.nih.gov/Traces/sra/sra.cgi?view=software) and aligns them to the genome using HISAT2 (Kim et al. 2019). It is also possible to use a combination of these formats when using more than one library.

Figure 2.

Average precision and sensitivity of gene predictions made by BRAKER1, BRAKER2, TSEBRA, GeneMark-ETP, and BRAKER3 for the genomes of 11 different species (listed in Supplemental Table S1). Inputs were the genomic sequences, short-read RNA-seq libraries, and protein databases (order excluded).

Figure 3.

Gene-level precision and sensitivity of gene predictions made by BRAKER1, BRAKER2, TSEBRA, GeneMark-ETP, and BRAKER3 for the genomes of 11 different species: well-annotated and compact genomes (first and second row), well-annotated and large genomes (third row), other genomes (fourth row). The fourth column shows the average for each group. Inputs were the genomic sequences, short-read RNA-seq libraries, and protein databases (order excluded).

Figure 4.

Lowly, medium, and highly expressed transcripts are in the first, second, and third terciles of expression levels, respectively.

Figure 5.

Average precision and sensitivity of gene predictions made by MAKER2, Funannotate, and BRAKER3 for a subset of eight species (excluding the mouse, spider, and fish genomes). Inputs were the genomic sequences, short-read RNA-seq libraries, and protein databases (close relatives included). The accuracy of MAKER2 reported here can be regarded as an upper limit of what can be expected when annotating a previously unannotated genome (see “Experiments” section).

Figure 6.

The execution time of BRAKER3. The time required for aligning the RNA-seq to the genome and thus producing the BAM input files is not included.