Survey
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
Download Unsupervised RNA-Seq-based genome annotation with GeneMark
Document related concepts
no text concepts found
Transcript
Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke RNA-Seq & gene prediction GeneMark-ET Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS January 11th 2015 AUGUSTUS BRAKER1 Pipeline Data sets Results Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke Corresponding author: [email protected] 1.1 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke Contents 1 RNA-Seq & gene prediction RNA-Seq & gene prediction 2 GeneMark-ET GeneMark-ET AUGUSTUS BRAKER1 Pipeline 3 AUGUSTUS Data sets Results 4 BRAKER1 Pipeline Data sets Results 1.2 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS From RNA-Seq to Genes Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke RNA-Seq & gene prediction GeneMark-ET AUGUSTUS BRAKER1 Pipeline Data sets Results Approaches: • de novo transcript assembly -> mapping -> gene prediction • mapping -> genome-guided assembly -> gene prediction • mapping -> gene prediction 1.3 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS GeneMark-ET uses RNA-Seq for Training Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke RNA-Seq & gene prediction GeneMark-ET AUGUSTUS BRAKER1 Pipeline Data sets Results • employs unsupervised training • includes in training introns and exons anchored by mapped RNA-Seq reads • does not require RNA-Seq reads assembly • does not use RNA-Seq information in the prediction step 1.4 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke GeneMark-ET uses RNA-Seq for Training Anchors from RNA-Seq for training RNA-Seq & gene prediction GeneMark-ET AUGUSTUS BRAKER1 Pipeline Data sets Results • employs unsupervised training • includes in training introns and exons anchored by mapped RNA-Seq reads • does not require RNA-Seq reads assembly • does not use RNA-Seq information in the prediction step 1.4 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke AUGUSTUS uses RNA-Seq for Prediction Introns predicted by RNASeq read alignment RNA-Seq & gene prediction Genome GeneMark-ET AUGUSTUS gene predictions with “hints” from RNA-Seq AUGUSTUS BRAKER1 Pipeline Data sets Results • requires “prior data” for training • uses intron information from RNA-seq for prediction • no RNA-Seq assembly required 1.5 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke RNA-Seq & gene prediction GeneMark-ET AUGUSTUS BRAKER1 Pipeline Data sets Results 1.6 PAG 2014... Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS PAG 2014... Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke Our intention was to create a eukaryotic gene prediction tool that RNA-Seq & gene prediction GeneMark-ET AUGUSTUS • trains automatically BRAKER1 • improves state-of-the-art gene prediction accuracy Pipeline • uses RNA-Seq for training and prediction (as Data sets Results unassembled reads) • is easy to use 1.6 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke RNA-Seq & gene prediction GeneMark-ET AUGUSTUS BRAKER1 Pipeline Data sets Results 1.7 BRAKER1 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke BRAKER1 Running BRAKER1 braker.pl [OPTIONS] -genome=genome.fa -bam=rnaseq.bam RNA-Seq & gene prediction GeneMark-ET AUGUSTUS ∼ 1 day for fly on 1 CPU BRAKER1 Pipeline Data sets Results Results • BRAKER1-GeneMark-ET gene predictions • BRAKER1-AUGUSTUS gene predictions 1.7 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke Data sets for accuracy evaluation Model organisms and Illumina paired end libraries • Drosophila melanogaster (flybase.org) • genome and reference annotation version R5 • RGASP RNA-Seq libraries RNA-Seq & gene prediction • Arabidopsis thaliana (arabidopsis.org) GeneMark-ET • genome and reference annotation version TAIR 10 • SRR934391 AUGUSTUS BRAKER1 Pipeline • Caenorhabditis elegans (wormbase.org) Data sets Results • genome and reference annotation version WS240 • RGASP RNA-Seq library • Schizosaccharomyces pombe (pombase.org) • genome and reference annotation version ASM294v2.23 • SRR097898, SRR097899, SRR097900, SRR097902, SRR097903, SRR097905, SRR097906, SRR097907, SRR097908, SRR097909, SRR097912, SRR097915, SRR097917, SRR097921, SRR097922, SRR097925, SRR402833 1.8 Accuracy of BRAKER1 C. elegans Data sets ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● BRAKER1− GeneMark−ET 1.9 ● Gene Sensitivity Gene Specificity Transcript Sensitivity Transcript Specificity Exon Sensitivity Exon Specificity BRAKER1− AUGUSTUS Pipeline Results ●● ● BRAKER1− GeneMark−ET BRAKER1 ● BRAKER1− AUGUSTUS % AUGUSTUS S. pombe ● ● ● ● A. thaliana BRAKER1− GeneMark−ET GeneMark-ET 86 81 76 71 66 61 56 51 46 41 36 31 BRAKER1− AUGUSTUS RNA-Seq & gene prediction D. melanogaster BRAKER1− GeneMark−ET Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke BRAKER1− AUGUSTUS Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke Comparing BRAKER1 to... MAKER21 ? Maker2 • generates “training genes” from assembled RNA-Seq • uses GeneMark-ES, AUGUSTUS, SNAP RNA-Seq & gene prediction • integrates RNA-Seq evidence (assembled and reads) into GeneMark-ET gene prediction AUGUSTUS BRAKER1 Pipeline How we use MAKER2 Data sets Results • no protein database • keep_preds=1 • include Cufflinks transcripts & read alignments • MAKER2 masks repeats 1.10 1 Following the tutorial at following tutorial at http://weatherby.genetics.utah.edu/MAKER/wiki/index.php/MAKER_Tutorial_ for_GMOD_Online_Training_2014 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Comparing BRAKER1 to MAKER2 Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke D. melanogaster C. elegans 38 ● 33 RNA-Seq & gene prediction BRAKER1 − MAKER2 ● GeneMark-ET AUGUSTUS BRAKER1 Pipeline Data sets Results 28 ● 23 18 13 ● ● ● ● ● ● ●● S. pombe ●● ● ● ● ● ● 8 3 A. thaliana Gene Sensitivity Gene Specificity Transcript Sensitivity Transcript Specificity Exon Sensitivity Exon Specificity ● ●●● ● ●● ●● −2 1.11 BRAKER1− AUGUSTUS BRAKER1− GeneMark−ET BRAKER1− AUGUSTUS BRAKER1− GeneMark−ET BRAKER1− AUGUSTUS BRAKER1− GeneMark−ET BRAKER1− AUGUSTUS BRAKER1− GeneMark−ET −7 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Future Work Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke RNA-Seq & gene prediction • integration of protein information GeneMark-ET AUGUSTUS • further optimization of BRAKER1 parameters BRAKER1 Pipeline • UTR training & integration of RNA-Seq coverage Data sets information Results 1.12 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke BRAKER1 is available for download at RNA-Seq & gene prediction http://bioinf.uni-greifswald.de GeneMark-ET AUGUSTUS BRAKER1 and Pipeline Data sets Results http://exon.gatech.edu 1.13 Unsupervised RNA-Seq-based genome annotation with GeneMark-ET & AUGUSTUS Acknowledgements Simone Lange, Katharina J. Hoff, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke RNA-Seq & gene prediction GeneMark-ET Simone Lange, Alexandre Lomsadze, Mark Borodovsky, Mario Stanke AUGUSTUS BRAKER1 Pipeline Data sets Results 1.14
Related documents