MAAB: A Motif and Amino-Acid Bias Pipeline to Identify Both Intrinsically Disordered and Repetitive Hydroxyproline-Rich Glycoproteins (HRGPs) in 1000 Plant Transcriptomes

Hydroxyproline-rich glycoproteins (HRGPs) are found in protein-rich cell walls of green algae and cellulose-rich walls of higher plants. HRGPs fall into three general classes; arabinogalactan-proteins (AGPs), extensins (EXTs) and proline-rich proteins (PRPs), spanning a continuum from highly glycosylated AGPs to minimally glycosylated PRPs. HRGPs represent a significant bioinformatics challenge as their core proteins have strong amino acid biases but relatively low sequence similarity. AGP core proteins are intrinsically disordered whereas extensin and PRP core proteins are repetitive. The recent transcriptome sequencing of over 1000 plant species in the 1000 plants (1KP) project (www.onekp.com) provides an opportunity to study the evolution of HRGPs at major evolutionary milestones. Building on previous work, we developed a motif & amino acid bias (MAAB) pipeline to classify HRGPs into more than twenty descriptive classes. Validation of the approach was done using available genomic resources [www.phytozome.net]. When applied to the 1KP dataset, improved recall of repetitive HRGPs was observed using multiple-K methodology, involving re-assembly of over 1200 transcriptome samples using four kmer-sizes (39, 49, 59, 69). This approach facilitated the identification of evolutionary trends across the HRGP classes. For example, GPI-anchored AGPs are found in some, but not all green algal lineages whereas cross-linking extensins and AG-peptides first appear in bryophytes, the first land plants. The transition to land was accompanied by an increased number of GPI-anchored AGPs but a decrease in the number of non-GPI-anchored AGPs. Selected results will be available from http://onekp.plantcell.unimelb.edu.au:8080/gwtphylo/ once published.