Perspect Ive Why We Need More Algal Genomes

In the current post-genomics world, a relevant question on the minds of many phycologists might be: do we really need more algal genomes or, should we stop and focus on the hard job of developing genetic tools and other resources for already sequenced taxa? This question has, in our opinion, a clear answer: we need to do both. Here, we focus on the genome sequencing side and discuss the following reasons why we think algal (and related heterotrophic protist) genome sequencing should remain a focus of phycological research: (1) transcriptomes that aim to create gene inventories or study gene expression differences (primarily Illumina RNAseq data), although cheap to produce and relatively easy to analyze, may not be sufficient for in-depth study of genomes, (2) much of natural biodiversity is still unstudied, necessitating approaches such as single cell genomics (SCG) that, although still challenging when applied to algae, can sample taxa isolated directly from the environment, (3) horizontal gene transfer (HGT) in algae is no longer controversial, but rather a major contributor to the evolution of photosynthetic lineages, and its study benefits greatly from completed (or draft) genomes, and (4) epigenetics and genome evolution among populations are best studied using assembled genome data. The power of RNAseq data to support gene discovery and, in particular, gene expression differences is clearly very high and has been exploited by many groups including the large-scale Marine Microbial Eukaryote Transcriptome Sequencing Project (MMETSP; Keeling et al. 2014; http://marinemicroeukaryotes.org/) and the 1,000 Plants (1KP; http://onekp.com/project.html) initiative and an array of more focused studies. These data provide significant opportunities for phycologists and are (still) the most appropriate approach for groups such as dinoflagellates that have enormous nuclear genomes that may be several times the size in humans (e.g., Hou and Lin 2009). Nonetheless, the benefits of RNAseq data can be greatly improved with the availability of reference genomes to build more accurate and complete gene models and to discriminate between nuclear “host” and contaminant genes. Whole genome analysis and annotation also provide information about gene structure (e.g., intron number and distribution; Nakamura et al. 2013), transposons (Read et al. 2013), and gene synteny (Blanc et al. 2012) that is invaluable for interpreting molecular evolution. In addition, sole use of transcriptome sequences may lead to assembly artifacts resulting in inflated gene counts and false contigs. This is partly explained by the significant coverage variation in assembled cDNA data that reflects gene expression differences across coding regions. This aspect, when compounded with sequencing errors in “over sequenced” areas can result in a poor or fragmented assembly (for discussion, see Martin and Wang 2011). In contrast, gene models predicted using complementary RNAseq and genome data generally provide a more complete and accurate inventory of nuclear genes, in particular, in identifying gene start and stop sites. Work with Porphyridium purpureum (Rhodophyta) showed that high quality RNAseq data alone identified 36,167 unique assembled contigs (N50 = 1,298nt), whereas only 8,355 genes were predicted in the genome. No obvious evidence of alternative splicing was found as an explanation for the high number of RNAseq-derived contigs in this red alga (Bhattacharya et al. 2013). Similarly, Shoguchi et al. (2013) assembled 63,104 unique RNAseq-derived contigs (N50 = 1,586nt) in a study of the coral symbiont Symbiodinium minutum (Dinophyceae), a genome that encodes ~42K genes. Therefore, genome data and the resulting gene models are the fundamental unit of study for many scientists and are used to provide templates for synthetic gene construction (Rockwell et al. 2014), to understand protein targeting (i.e., the N-terminus is needed for this purpose; Emanuelsson et al. 2000), to determine accurately gene family size (e.g., pherophorins in Volvox carteri; Prochnik et al. 2010), to have a reliable reference to map cDNA transcripts Received 22 August 2014. Accepted 10 November 2014. Author for correspondence: e-mail [email protected]. Editorial Responsibility: M. Graham (Managing Editor) J. Phycol. 51, 1–5 (2015) © 2014 Phycological Society of America DOI: 10.1111/jpy.12267