Does read length really matter when doing a de novo assembly of BACs from highly repetitive plant genomes, like that of barley? Will paired end sequencing improve the assembly considerably? Is it essential to barcode each BAC clone before sequencing?
These are the main questions of a study from Taudien et al, published in September 2011, answered by comparing assembly data derived from different sequencing data sets.
To investigate the effect of the read length the authors compared the assembly quality from 4 BAC clones with either FLX or with Titanium chemistry reads at equal sequencing depth. Even though the average read length differed by only 40 bp (223 vs 263 bp) the assemblies with the longer reads were considerably better. They contained fewer missassemblies and reduced number of gaps. According to the authors this is mainly due to the broader read distribution of the Titanium chemistry, where a fraction of reads displayed a read length of >600 bp.
These findings are in accordance with our experience from BAC assemblies. Having performed several BAC sequencing projects with GS FLX+ yet (average read length of 550-650 bp and some reads up to 1000 bp), we can report here that FLX+ read length has the potential to even more improve the BAC assembly quality.
Please have a look at the research article for further information.