transcriptomics. The high throughput and low
cost of NGS sequencing technologies also
advance the sequencing of many crops with a
genome size of less than 500 Mb, including
cucumber (Huang et al. 2009), watermelon (Guo
et al. 2013), wild strawberry (Fragaria vesca)
(Kang et al. 2013), date palm (Al-Dous et al.
2011), and papaya (Ming et al. 2008), providing
invaluable genomic resources for the breeding of
vegetables and fruit trees. The increase of data
volume and the improvement of accuracy make
NGS a great advantage in sequencing most
duckweed genomes, especially the genomes with
small genome size and low complexity. Before
the initiation of duckweed genome sequencing
(Wang et al. 2011), a series of genome sizes of
duckweed species were determined by flow
cytometry (FCM). The species with small genome sizes have been selected and sequenced by
the next-generation sequencing technologies.
Spirodela polyrhiza 7498 (2n = 40; estimated
158 Mb) was sequenced by Sanger and 454
Roche platforms. The length of contig N50
arrived to 17.8 kb. With the integration of physical map and BAC end sequencing, the scaffolds
were generated with the N50 of 7.6 Mb (Wang
et al. 2014). Spirodela polyrhiza 9509 (2n = 40;
estimated 158 Mb) was sequenced by Illumina
HiSeq 2000 platform. A total of two paired-end
(PE) and three mate-pair (MP) sequencing
libraries were constructed from its genomic DNA,
with insert sizes of 180 bp, 500 bp, 2 kb, 5 kb,
and 20 kb. The 180 bp and 500 bp libraries were
used for de novo assembling into contigs, and the
2 kb, 5 kb, and 20-kb libraries were targeted for
constructing scaffolds, resulting in a size of 145.8
Mb assembled genome with a scaffold N50 of
4.3 Mb and a contig N50 of 19 kb (Michael et al.
2017). The Lemna minor 5500 (2n = 40; estimated 481 Mb), as the Lemna ancestor genome,
was sequenced by Illumina platform. Two
paired-end libraries were created including a
HiSeq library (2 Â 100 bp) and a MiSeq library
with longer short reads (2 Â 300 bp). The draft
genome was assembled into 472.1 Mb containing
46,047 scaffolds with an N50 length of 23.8 kb
and a contig N50 length of 20.9 kb (2015).
Compared with Spirodela polyrhiza, the
assembly of Lemna minor 5500 was more fragmented and contained more gaps. One reason was
that mate-pair (MP) libraries with large DNA
insertions were missing, which could help bridge
contigs into scaffolds. Another was that 61.5% of
the genome size (481 Mb) was repetitive
sequences in Lemna minor 5500 that was challenging to be determined. In contrast, there was
only 15.79% of repeat elements in the Spirodela
polyrhiza genome (158 Mb). The repeat content
could explain 94.5% of the genome size difference between Lemna minor and Spirodela polyrhiza (Van Hoeck et al. 2015).
Unlike mammalian genomes, plant genomes
are abundant of transposons, leading to a huge
variation in genome size. The smallest genome,
Genlisea tuberosa, is only 61 Mb in size that
basically maintains the essential genes. The wheat
genome has a large genome size of 17 Gb, of
which 90% is a repeat sequence (Michael and
VanBuren 2015). The loblolly pine genome even
reached 22 Gb that is the largest sequenced genome until recently (Zimin et al. 2014). However,
the short read from Illumina sequencing
(*150 bp) limits its applications since the
sequence reads cannot fully span the repeat
regions and pose a serious problem for genome
assembly. The similarity of repeat reads breaks
the contiguity of genome into fragmented contigs,
resulting in incomplete genome assembly. The
unassembled sequences may confound the biological significance due to the missing of a complete gene, partial regulatory elements, impaired
centromeres, and telomeres (Li et al. 2018).
Duckweeds genome sizes vary enormously,
ranging from 150 Mb in Spirodela to 1881 Mb in
Wolffia, a total 13-fold change (Wang et al. 2011).
The difference of their genome sizes is mainly
caused by repetitive sequences, especially long
terminal repeats (LTR) which have a length of
4–16 kb and huge amount of copies (Kumar and
Bennetzen 1999; Phillippy 2017). Only the long
reads that expand over the transposons can make
their uniqueness so that they can determine where
the repeats belong to. Given the availability of
third-generation sequencing (TGS), scientists
have an extraordinary opportunity to crack the
complex duckweed genomes.
68
X. Xiang and C. Li
cost of NGS sequencing technologies also
advance the sequencing of many crops with a
genome size of less than 500 Mb, including
cucumber (Huang et al. 2009), watermelon (Guo
et al. 2013), wild strawberry (Fragaria vesca)
(Kang et al. 2013), date palm (Al-Dous et al.
2011), and papaya (Ming et al. 2008), providing
invaluable genomic resources for the breeding of
vegetables and fruit trees. The increase of data
volume and the improvement of accuracy make
NGS a great advantage in sequencing most
duckweed genomes, especially the genomes with
small genome size and low complexity. Before
the initiation of duckweed genome sequencing
(Wang et al. 2011), a series of genome sizes of
duckweed species were determined by flow
cytometry (FCM). The species with small genome sizes have been selected and sequenced by
the next-generation sequencing technologies.
Spirodela polyrhiza 7498 (2n = 40; estimated
158 Mb) was sequenced by Sanger and 454
Roche platforms. The length of contig N50
arrived to 17.8 kb. With the integration of physical map and BAC end sequencing, the scaffolds
were generated with the N50 of 7.6 Mb (Wang
et al. 2014). Spirodela polyrhiza 9509 (2n = 40;
estimated 158 Mb) was sequenced by Illumina
HiSeq 2000 platform. A total of two paired-end
(PE) and three mate-pair (MP) sequencing
libraries were constructed from its genomic DNA,
with insert sizes of 180 bp, 500 bp, 2 kb, 5 kb,
and 20 kb. The 180 bp and 500 bp libraries were
used for de novo assembling into contigs, and the
2 kb, 5 kb, and 20-kb libraries were targeted for
constructing scaffolds, resulting in a size of 145.8
Mb assembled genome with a scaffold N50 of
4.3 Mb and a contig N50 of 19 kb (Michael et al.
2017). The Lemna minor 5500 (2n = 40; estimated 481 Mb), as the Lemna ancestor genome,
was sequenced by Illumina platform. Two
paired-end libraries were created including a
HiSeq library (2 Â 100 bp) and a MiSeq library
with longer short reads (2 Â 300 bp). The draft
genome was assembled into 472.1 Mb containing
46,047 scaffolds with an N50 length of 23.8 kb
and a contig N50 length of 20.9 kb (2015).
Compared with Spirodela polyrhiza, the
assembly of Lemna minor 5500 was more fragmented and contained more gaps. One reason was
that mate-pair (MP) libraries with large DNA
insertions were missing, which could help bridge
contigs into scaffolds. Another was that 61.5% of
the genome size (481 Mb) was repetitive
sequences in Lemna minor 5500 that was challenging to be determined. In contrast, there was
only 15.79% of repeat elements in the Spirodela
polyrhiza genome (158 Mb). The repeat content
could explain 94.5% of the genome size difference between Lemna minor and Spirodela polyrhiza (Van Hoeck et al. 2015).
Unlike mammalian genomes, plant genomes
are abundant of transposons, leading to a huge
variation in genome size. The smallest genome,
Genlisea tuberosa, is only 61 Mb in size that
basically maintains the essential genes. The wheat
genome has a large genome size of 17 Gb, of
which 90% is a repeat sequence (Michael and
VanBuren 2015). The loblolly pine genome even
reached 22 Gb that is the largest sequenced genome until recently (Zimin et al. 2014). However,
the short read from Illumina sequencing
(*150 bp) limits its applications since the
sequence reads cannot fully span the repeat
regions and pose a serious problem for genome
assembly. The similarity of repeat reads breaks
the contiguity of genome into fragmented contigs,
resulting in incomplete genome assembly. The
unassembled sequences may confound the biological significance due to the missing of a complete gene, partial regulatory elements, impaired
centromeres, and telomeres (Li et al. 2018).
Duckweeds genome sizes vary enormously,
ranging from 150 Mb in Spirodela to 1881 Mb in
Wolffia, a total 13-fold change (Wang et al. 2011).
The difference of their genome sizes is mainly
caused by repetitive sequences, especially long
terminal repeats (LTR) which have a length of
4–16 kb and huge amount of copies (Kumar and
Bennetzen 1999; Phillippy 2017). Only the long
reads that expand over the transposons can make
their uniqueness so that they can determine where
the repeats belong to. Given the availability of
third-generation sequencing (TGS), scientists
have an extraordinary opportunity to crack the
complex duckweed genomes.
68
X. Xiang and C. Li