Maritime pine (Pinus pinaster) de novo transcriptome assembly

Rationale Fast increase of world population leads to a correlative increase on wood demand (heating, pulp and paper, building). Consequently, the cultivated forest has a leading role to play to fill this growing demand. Developing an abundant wood ressource calls for improvement of sylvicultural practices as well as deployment of genetically improved trees varieties in the frame of a sustainable forestry. Maintaining french forest competitivity will depend on the integration of new technologies in breeding programs, especially for intensively cultivated forest species such as maritime pine (Pinus pinaster). Second generation sequencing platforms provide tremendous and relatively cheap amounts of data to biologists. However, the main characteristic of this data is to be extremely puzzled with about 20 millions of short sequences averaging 100bp length per sample. Aligning such data along a reference genome is a routine task nowadays. But in the case of maritime pine, de novo assembling of a transcriptome (i.e. without reference genome) from the 300 millions of sequences available, mixing 1% of Sanger and 454 long sequences and 99% of such short reads, is quite challenging and calls for delicate algorithmic approaches (Wall et al. 2009). Nevertheless, applying short sequences to transcriptomics is already possible (Wheat 2010) using for instance informations provided by orthologous genes (Birzele et al. 2010) and promises to be an acceptable experimental solution for many non-model agronomical species. Many different algorithms for transcriptome assembly are available (Miller et al. 2010, Surget-Groba & Montoya-Burgos 2010). Kumar & Blaxter (2010) compared CAP3, MIRA3, Newbler, SeqMan and CLC for the de novo assembly of Litomosoides sigmodontis nematode transcriptome. In the latter paper, an interesting idea was to fuse different assemblies to maximise cooperation of different algorithms in the final assembly. Aims In this poster, we will describe a preliminary assembly of 454 sequences by TGICL (Pertea et al. 2003) and we propose to study the usability of several tools of transcriptome assembly, such as MIRA3 (Chevreux et al. 2004), Velvet (Zerbino & Birney 2008), ABySS (Birol et al. 2009) and SOAPdeNovo (Li et al. 2010), integrating massive short reads, and define an acceptable method to apply to maritime pine transcriptomics