Sequencing strategy: Low-coverage shotgun sequencing of genomic DNA
-------------------------------------------------------------------

The resulting data of low-coverage shotgun sequencing of genomic DNA (gDNA), aka genome skimming, is the primary data used by ``ORG.asm``. If we hypophethize that the organelle genomes represent several percent of the total gDNA (organellar genomes can be present in more than 1000 copies in a single cell), even with a modest depth of sequencing of the nuclear genome (around 1x coverage), on can hope to get more than 100x coverage for the organelle genomes and repeated regions (such as rDNA clusters). This allows the reconstruction of organelle genomes and repeated regions for up to 48 samples loaded in the same HiSeq 2500 lane. 


For example, Consider that you sequence 3.10e6 pair-end reads -> 6.10e6 reads of 100bp

==================== ============ ============
Organelle            Chloroplast  Mitochondria
==================== ============ ============
Belonging organelle	 5%           0.5%
Effective reads	     300,000      30,000
Base pairs           30.10e6      3.10e6
Genome size          150kb        16Kb
Sequencing depth     200X         187X
==================== ============ ============


This can be further observed using the k-mer frequency spectrum of a plant genome low-coverage shotgun sequencing.
The spectrum shows particular a shape with a high number of non-frequent kmer and a bimodal shape at intermediate and high frequency. 

This shape can be explained by the mix of the low coverage sequencing the high coverage sequecing of the chloroplastic genome. Indeed the nuclear genome sequencing is responsible for a large number of unique or nearly unique k-mer and the high coverage sequencing of the chloroplastic genome translate into the bimodal distribution of moderatly and highly encountered k-mers, the bimodal distribution being due to the large duplicated region (Inverted Repeat) typical of the chloroplastic genome.

.. figure:: ./Kmer-histogram.*
  :align: center
  :figwidth: 80 %
  :width: 500
