SL2.50_REPET.gff3
Annotation performed with the REPET package v2.2 (Maumus & Quesneville 2014). A library of consensus sequences was built de novo using the TEdenovo pipeline using a subset of the SL2.50 assembly (considering only the contigs of length above 100 kb, 334 Mb total). Classification of the consensus sequences was performed using PASTEC (Hoede et al.) followed by semi-manual curation. Target is the name of the consensus used to annotate a feature followed by the positions on the consensus.

SL2.50_RSRM.gff3
RepeatScout (v1.0.5, Price et al. 2005) was launched on a subset of the SL2.50 assembly (considering only the contigs of length below 100 kb, 403 Mb total) with the option -stopafter 500. Redundancy was removed from the RepeatScout library following thresholds of identity greater than 95% and length greater than 98%. Classification of the consensus sequences was performed using PASTEC (Hoede et al.) followed by semi-manual curation. Target is the name of the consensus used to annotate a feature followed by the positions on the consensus.

SL2.50_Tallymer.gff3
TallymerPipe.py from the REPET package was used to launch Tallymer (Kurtz et al. 2008)on the whole assembly with parameters of word size = 16 and minimum occurences = 10. The annotations included within positions detected by Tandem Repeats Finder were discarded. The positions not included in the REPET and RepeatScout annotations were computed and the remaining fragments of length below 30 bp were discarded. 

SL2.50_TRF.gff3
Tandem Repeats Finder (v4.00, Benson 1999) was launched on the SL2.50 assembly with the following set of parameters: 2.7.7.80.10.50.500. 

The paper is under review and this document will be updated once it is online.
Comprehensive Repeatome Annotation Reveals Strong Potential Impact of Repetitive Elements on Tomato Ripening.
Ophélie Jouffroy; Surya Saha; Lukas Mueller; Hadi Quesneville; Florian Maumus (submitted)