158
Xenopus
2012, these two teams reached an agreement to collaborate;
as the project progressed, other researchers joined to form an
international consortium.
The genome sequencing of X. laevis was challenging not
only because of its estimated large genome size (3 Gbp, nearly
twice as much of X. tropicalis and comparable to human, with
a total of 36 chromosomes), but the biggest obstacle was to
distinguish the highly conserved pairs of homeologous genes
and to correctly assemble the whole genome. To overcome the
problem with heterozygosity, in which allelic variations potentially disturb homeolog-specifc assembly, the highly inbred
“J-strain” was chosen as the source of DNA for sequencing.
The J-strain was made by inbreeding frogs by C. Katagiri, S.
Tochinai, and their colleagues in Hokkaido University, Japan
(see Extended Data Figure 1a in Session et al., 2016 , for the
history of the J-strain). Originally, this strain was made as a
model animal for immunological analyses and has been used
in several studies ( Izutsu and Yoshizato, 1993 ; Robert and
Ohta, 2009 ). At the time the whole genome sequencing project
started, inbreeding had been conducted for more than 30 generations, and even before that, long-term immunological rejection did not occur in skin transplantation assays ( Izutsu and
Yoshizato, 1993 ). Therefore, polymorphic sequences could
most likely be attributed to the difference between homeologs.
As X. laevis has a female heterozygous (ZW) sex determination system (Chang and Witschi, 1956), DNA from
females of the J-strain was shotgun and mate pair sequenced
with Illumina. Contigs and scaffolds were constructed incorporating end-sequencing of BAC and fosmid clones. These
sequences were further assembled to chromosome-scale
using several methods: FISH analyses with BACs and in
vivo and in vitro chromatin conformation capture methods
(HiC and the Chicago method, respectively) (LiebermanAiden et al., 2009; Putnam et al., 2016). FISH analyses identifed several chimeric scaffolds, which were then separated
by estimating the fusion sites and reassembled. Chimeric
sequences were also identifed by randomly and arbitrarily
selecting 987 genes and testing the integrity of their gene
structures. In the case of tandemly repeated homologous
gene clusters, such as Hox clusters and mix/bix gene clusters and regions whose sequence could not be fully determined by automatic assembly of shotgun sequences, they
were flled by manually selecting the corresponding BAC
and fosmid clones and determining their entire sequences.
Finally, the ver. 9.1 assembly contained chromosome-level
sequences for each of the 18 chromosomes. Through meticulous verifcation of the assembly, feedback, and re-assembly,
the genome sequence of X. laevis became probably one of
the most reliable genomes of its size. The current version
(as of April 2021) is ver. 10.1, which was reassembled using
PacBio long read sequences (Genbank GCA_017654675.1).
10.2.3. COMPARISON BETWEEN X. TROPICALIS AND
X. LAEVIS GENOMES AND CHROMOSOMES
Before determining the whole genome sequence of X. laevis,
the correspondence between the chromosomes of X. tropicalis
(XTR) and X. laevis (XLA) was identifed by FISH analysis
using cDNAs of X. laevis ( Uno et al., 2013 ), as well as by the
detailed chromosome map of X. laevis constructed with the
assignment of BAC clones ( Matsuda et al., 2015 ; Session et
al., 2016 ). FISH analyses with cDNA probes determined the
pairs of homeologous chromosomes of X. laevis, since homeologous chromosomes harbored basically the same genes.
As mentioned before, X. tropicalis is diploid and has 20
chromosomes (2n = 2x = 20); the chromosome pairs are
numbered as XTR1 to XTR10 ( Khokha et al., 2009 ). The
allotetraploid X. laevis has 36 chromosomes (2n = 4x = 36),
consisting of nine pairs of homoeologous chromosomes.
Of them, eight pairs each corresponded to one X. tropicalis
chromosome (XTR1–XTR8), whereas the remaining XLA
chromosomes corresponded to a fusion between XTR9 and
XTR10 ( Uno et al., 2013 ). Based on the correspondence with
X. tropicalis chromosomes, the nine homoeologous chromosome sets of X. laevis were renumbered as XLA1, XLA2,
XLA3, and so on. As one of the homeologous chromosomes
is longer than the other, according to relative lengths measured with the karyotypes (Matsuda et al., 2015), the longer chromosomes were suffxed by adding the letter L (for
“long”) and the shorter chromosomes with S (for “short”),
making XLA1L, XLA1S, XLA2L, and so on (see Figure
10.1B for the nomenclature and relationships of the chromosomes). The chromosomes homologous to XTR9 and
XTR10 were named XLA9_10L and XLA9_10S to ref ect
their fused status, or simply XLA9L and XLA9S. As stated
in Matsuda et al. ( 2015 ), cytogeneticists may prefer the simple nomenclature, but XLA9_10L and XLA9_10S are more
convenient to perform direct genome and chromosome comparisons between X. laevis and X. tropicalis.
Chromosome fusion sites in XLA9_10 were identif ed
by a comparison of the synteny of genes of XLA9_10L
and XLA9_10S and those at the ends of XTR9 and XTR10
( Session et al., 2016 ). Since the fusion regions in XLA9_10L
and XLA9_10S were identical and the basic number of
chromosomes in the family Pipidae (including the genus
Xenopus) was n = 10, the chromosome fusion was suggested
to have occurred after the divergence from X. tropicalis and
before the speciation of the two diploid ancestor species of X.
laevis. Analyses showed that two chromosomes corresponding to XTR9 and XTR10 in an ancestral frog fused tandemly
(without any obvious gene loss) and the centromere was
repositioned during karyotype evolution (see Extended Data
Figure 2 in Session et al., 2016 ).
10.2.4. THE REAL STORY OF THE IDENTIFICATION OF
L AND S SUBGENOMES AND CHROMOSOMES
X. laevis homeologous chromosomes were designated with
postfxes L and S just according to the difference in lengths
of chromosomes, as mentioned previously. The description
of chromosomes and subgenomes in Session et al. (2016 )
may give the impression that the subgenomes coincidentally
corresponded to the sets of the L (longer) and S (shorter)
chromosomes of each homeologous pair, but in fact, we had
identifed the subgenomes before naming the chromosomes.
Here is the real story of how L and S were adopted.
Xenopus
2012, these two teams reached an agreement to collaborate;
as the project progressed, other researchers joined to form an
international consortium.
The genome sequencing of X. laevis was challenging not
only because of its estimated large genome size (3 Gbp, nearly
twice as much of X. tropicalis and comparable to human, with
a total of 36 chromosomes), but the biggest obstacle was to
distinguish the highly conserved pairs of homeologous genes
and to correctly assemble the whole genome. To overcome the
problem with heterozygosity, in which allelic variations potentially disturb homeolog-specifc assembly, the highly inbred
“J-strain” was chosen as the source of DNA for sequencing.
The J-strain was made by inbreeding frogs by C. Katagiri, S.
Tochinai, and their colleagues in Hokkaido University, Japan
(see Extended Data Figure 1a in Session et al., 2016 , for the
history of the J-strain). Originally, this strain was made as a
model animal for immunological analyses and has been used
in several studies ( Izutsu and Yoshizato, 1993 ; Robert and
Ohta, 2009 ). At the time the whole genome sequencing project
started, inbreeding had been conducted for more than 30 generations, and even before that, long-term immunological rejection did not occur in skin transplantation assays ( Izutsu and
Yoshizato, 1993 ). Therefore, polymorphic sequences could
most likely be attributed to the difference between homeologs.
As X. laevis has a female heterozygous (ZW) sex determination system (Chang and Witschi, 1956), DNA from
females of the J-strain was shotgun and mate pair sequenced
with Illumina. Contigs and scaffolds were constructed incorporating end-sequencing of BAC and fosmid clones. These
sequences were further assembled to chromosome-scale
using several methods: FISH analyses with BACs and in
vivo and in vitro chromatin conformation capture methods
(HiC and the Chicago method, respectively) (LiebermanAiden et al., 2009; Putnam et al., 2016). FISH analyses identifed several chimeric scaffolds, which were then separated
by estimating the fusion sites and reassembled. Chimeric
sequences were also identifed by randomly and arbitrarily
selecting 987 genes and testing the integrity of their gene
structures. In the case of tandemly repeated homologous
gene clusters, such as Hox clusters and mix/bix gene clusters and regions whose sequence could not be fully determined by automatic assembly of shotgun sequences, they
were flled by manually selecting the corresponding BAC
and fosmid clones and determining their entire sequences.
Finally, the ver. 9.1 assembly contained chromosome-level
sequences for each of the 18 chromosomes. Through meticulous verifcation of the assembly, feedback, and re-assembly,
the genome sequence of X. laevis became probably one of
the most reliable genomes of its size. The current version
(as of April 2021) is ver. 10.1, which was reassembled using
PacBio long read sequences (Genbank GCA_017654675.1).
10.2.3. COMPARISON BETWEEN X. TROPICALIS AND
X. LAEVIS GENOMES AND CHROMOSOMES
Before determining the whole genome sequence of X. laevis,
the correspondence between the chromosomes of X. tropicalis
(XTR) and X. laevis (XLA) was identifed by FISH analysis
using cDNAs of X. laevis ( Uno et al., 2013 ), as well as by the
detailed chromosome map of X. laevis constructed with the
assignment of BAC clones ( Matsuda et al., 2015 ; Session et
al., 2016 ). FISH analyses with cDNA probes determined the
pairs of homeologous chromosomes of X. laevis, since homeologous chromosomes harbored basically the same genes.
As mentioned before, X. tropicalis is diploid and has 20
chromosomes (2n = 2x = 20); the chromosome pairs are
numbered as XTR1 to XTR10 ( Khokha et al., 2009 ). The
allotetraploid X. laevis has 36 chromosomes (2n = 4x = 36),
consisting of nine pairs of homoeologous chromosomes.
Of them, eight pairs each corresponded to one X. tropicalis
chromosome (XTR1–XTR8), whereas the remaining XLA
chromosomes corresponded to a fusion between XTR9 and
XTR10 ( Uno et al., 2013 ). Based on the correspondence with
X. tropicalis chromosomes, the nine homoeologous chromosome sets of X. laevis were renumbered as XLA1, XLA2,
XLA3, and so on. As one of the homeologous chromosomes
is longer than the other, according to relative lengths measured with the karyotypes (Matsuda et al., 2015), the longer chromosomes were suffxed by adding the letter L (for
“long”) and the shorter chromosomes with S (for “short”),
making XLA1L, XLA1S, XLA2L, and so on (see Figure
10.1B for the nomenclature and relationships of the chromosomes). The chromosomes homologous to XTR9 and
XTR10 were named XLA9_10L and XLA9_10S to ref ect
their fused status, or simply XLA9L and XLA9S. As stated
in Matsuda et al. ( 2015 ), cytogeneticists may prefer the simple nomenclature, but XLA9_10L and XLA9_10S are more
convenient to perform direct genome and chromosome comparisons between X. laevis and X. tropicalis.
Chromosome fusion sites in XLA9_10 were identif ed
by a comparison of the synteny of genes of XLA9_10L
and XLA9_10S and those at the ends of XTR9 and XTR10
( Session et al., 2016 ). Since the fusion regions in XLA9_10L
and XLA9_10S were identical and the basic number of
chromosomes in the family Pipidae (including the genus
Xenopus) was n = 10, the chromosome fusion was suggested
to have occurred after the divergence from X. tropicalis and
before the speciation of the two diploid ancestor species of X.
laevis. Analyses showed that two chromosomes corresponding to XTR9 and XTR10 in an ancestral frog fused tandemly
(without any obvious gene loss) and the centromere was
repositioned during karyotype evolution (see Extended Data
Figure 2 in Session et al., 2016 ).
10.2.4. THE REAL STORY OF THE IDENTIFICATION OF
L AND S SUBGENOMES AND CHROMOSOMES
X. laevis homeologous chromosomes were designated with
postfxes L and S just according to the difference in lengths
of chromosomes, as mentioned previously. The description
of chromosomes and subgenomes in Session et al. (2016 )
may give the impression that the subgenomes coincidentally
corresponded to the sets of the L (longer) and S (shorter)
chromosomes of each homeologous pair, but in fact, we had
identifed the subgenomes before naming the chromosomes.
Here is the real story of how L and S were adopted.