This region is easy to assemble with Nanopore sequences from males as they have only a single copy of it.
Pairwise alignments to assemblies of the O9, O3, OSK42 and AOM-5 all support the following arrangement until S1. S66- is added after S1 after more detailed inspection (support in at least O3-3f.
XSR: S10+, S9+, S32+, S84-, S59+, S8-, S1+, S66-
Some assemblies have telomeric sequences at the beginning of their XSR contig before the matches with S10.
Inspection of the assembly graph suggest that the other end of the XSR contigs in Flye assemblies is followed by the rDNA region, which we know is at the beginning of the pseudo-autosomal region on the short arm of the pseudo-autosomal chromosome.
Attempts to incorporate smaller OSKA2016 scaffolds only leads to discrepancies of support from Flye assembly.
Thus, the XSR superscaffold above appears to be the best we can produce, and is properly oriented.
It appears that this regions accumulates variations at such a fast rate that O9 and O3 do not fully agree on how to rescaffold OSKA2016, and OSK42 and AOM individuals are clearly different.
Let's arbitrarly use O3-3f as a reference for superscaffolding now. This is the individual that was used in the OSKA2016 genome paper. O3-3f's contig_4 starts with a telomere sequence, suggesting it is properly oriented.
YSR: S63-, S71-, S68-, S48-, S55+, S42-, S65-, S87+, S86+, S78+, S110+, S62+, S39+, S26+, S47+, S64-
Orientation unknown, probably contains errors, and original scaffolds also probably contain misassemblies.
S29+, S49-, S33+, S80-, S119+, S93+, S43+, S128+, S83+, S57-, S56+, S45+, S117-, S88+, S92-, S21+, S12+, S79+, S89+, S53+, S54-, S11-, S74+, S25-
S13 matches telomeric contigs in assemblies and is colinear to S70 and S52.
- S34-, S94-, S2-, S18+, S73-, S23+, S52-, S70-, S13-
S31 matches a contig starting with a telomere, and belongs to the short arm.
Let's assume that chr1 starts with:
S31-, S51-, S38-, S85+, S61-, S44-, S109+, S37-
Then there is S6, with unkonwn orientation. Therefore S6+ was arbitrarly chosen.
S36+ S76+ S58+ are associated with S6 with unknown confidence except that they are part of this arm. Added in arbitrary position.
Also, S28 belongs to this arm, but is not well aligning to the Nanopore assemblies and I can not place it correctly. Added arbitrarly between S37 and S36.
chr1 short arm: S31-, S51-, S38-, S85+, S61-, S44-, S109+, S37-, S28+, S36+, S76, S58+, S6+
S3 matches a contig starting with a telomere. But it also crosses a centromere as it matches the left and right arm at the same time. Unless it is a missasembly.
There is good support that S5+ is followed by S7- but what is the orientation of the whole ?
Provisionally, let's use S6+ S3+ S5+ S7-
Multiple Nanopore assemblies have a contig that starts (when properly oriented) with a telomere and then matches S82.
chr2 short arm: S82-, S30+, S15-, S50-, S41+, S17+, S16-, S20+
Luckily, O33f's scaffold 2 is the whole long arm of Chr2. And S4's tail aligns to a contig ending with a telomere ! Dead end on the other side.
chr2 long arm: S60-, S14-, S90-, S22-, S35+, S118+, S46+, S102-, S72-, S104+, S40+, S27+, S75-, S24-, S19+, S101+, S4-
OSKA_V2_USAs='XSR.usa YSR.usa PAR.usa Chr1.usa Chr2.usa oskaV2.usa'
# No contig of OSKA2016.usa is not found in oskaV2
grep OSKA201 $OSKA_V2_USAs | sed -e 's/\[.*//g' -e 's/.*fasta//g' -e 's/$/(\\[::r])?$/' | grep -Evf - OSKA2016.usa
# No contig of OSKA2016.usa is not used in oskaV2
cat OSKA2016.usa | sed -e 's/\[.*//g' -e 's/.*fasta//g' -e 's/$/(\\[::r])?$/' | grep -Evf - $OSKA_V2_USAs
# Check that no contig is used twice !
grep fasta $OSKA_V2_USAs | wc -l # Should be 576
cat $OSKA_V2_USAs | grep fasta | sed 's/\[.*//'| sort | uniq --repeated # Should not return anything.
Genomes were masked for repeats using https://github.com/oist/LuscombeU_stlrepeatmask/tree/main v1.0.0
softmasked genomes are denoted by the suffix .softmasked.fa