Repeat masking

This data set is part of the Polistes dominula genome project, and provides details regarding the annotation and masking of transposable elements and other repetitive sequences in the P. dominula genome, as described in (Standage et al., Molecular Ecology, 2016). Included in this data set are the repeat annotations, the masked genome sequence, and documentation providing complete disclosure of the masking procedure.

Synopsis

The genome assembly was screened for known repetitive elements using RepeatMasker version open-4.0.5 and Repbase version 20140131. After masking repeats identified by RepeatMasker, the assembly was screened for additional repeats using Tallymer version 1.5.2. To discriminate bona fide repetitive elements from genes occurring in high copy number in the genome, all repeats identified by Tallymer were subjected to a BLASTX search against a database of Hexapod proteins. Any repeats with matches in the database and e-values < 1e-5 were discarded as probable high copy number genes, while the rest were used to mask the genome.

Data access

The unmasked P. dominula reference genome assembly is available for download at the DOI 10.6084/m9.figshare.1593098.

Procedure

First, we identified known repeats with RepeatMasker. By default, RepeatMasker produces soft-masked (lower-case) sequences, so we need to post-process the output to hard mask (N) the sequence.

NumProcs=16
GCContent=30.77
RepeatMasker -species insects -parallel $NumProcs -gc $GCContent \
             -frag 4000000 -lcambig -xsmall -gff \
             pdom-scaffolds-unmasked-r1.2.fa \
             > rm.log 2>&1
python lc2n.py < pdom-scaffolds-unmasked-r1.2.fa.masked \
               > pdom-rm-masked.fa

Next, we performed additional k-mer based screening for repetitive elements using Tallymer (procedure published by Dan Bolser).

gt suffixerator -v \
                -db $IDX \
                -indexname $IDX \
                -tis -suf -lcp -des -ssp -sds -dna \
                > suffixerator.log 2>&1

gt tallymer occratio -v \
                     -minmersize 10 \
                     -maxmersize 45 \
                     -output unique nonunique nonuniquemulti total relative \
                     -esa $IDX \
                     > pdom.occratio.10.45.dump

gt tallymer mkindex -v \
            -mersize 19 \
            -minocc 50 \
            -esa $IDX \
            -counts -pl \
            -indexname pdom.idx.19.50 \
            > mkindex.log 2>&1

gt tallymer search -v \
            -output qseqnum qpos counts \
            -tyr pdom.idx.19.50 \
            -q $IDX \
            > pdom.repeats.19.50.tmer \
            2> tallymer.search.log

tallymer2gff3.plx -k 19 -s $IDX \
                  pdom.repeats.19.50.tmer \
                  > pdom.repeats.19.50.gff3

gff2fasta.plx -s pdom-rm-masked.fa \
              -f pdom.repeats.19.50.gff3 \
              > pdom.repeats.19.50.fa

We then did a BLASTx search of repeats found by Tallymer vs known hexapod proteins, and parsed out those with hits using MuSeqBox.

curl 'http://www.uniprot.org/uniprot/?query=taxonomy%3a6960&force=yes&format=fasta' \
     > hexapoda.fa
makeblastdb -in hexapoda.fa -dbtype prot -parse_seqids
blastx -query pdom.repeats.19.50.fa -db hexapoda.fa \
       -num_alignments 10 -evalue 1e-5 -num_threads 64 \
       -out pdom.repeats.19.50.blastx \
       > pdom.repeats.19.50.log 2>&1

MuSeqBox -i pdom.repeats.19.50.blastx -L 100 \
    | cut -f 1 -d ' ' | sort | uniq \
    | perl -ne 'm/(PdomSCFr1.2-\d+)-\d+\/(\d+)-(\d+)/ and print "$1\t$2\t$3\n"' \
    > pdom.repeats.19.50.hexapodhits.txt

Finally, we masked the Tallymer repeats, excluding any that match Hexapod proteins as probably high-copy-number genes.

mask.pl pdom.repeats.19.50.gff3 \
        pdom.repeats.19.50.hexapodhits.txt \
        pdom-rm-masked.fa \
        > pdom-scaffolds-masked-r1.2.fa

---

This work is licensed under a Creative Commons Attribution 4.0 International License.