MILTEL - Method of Long-read TELomere detection

During development of MAC genomes in ciliates, chromosome breakage and addition of telomeric repeats to the newly formed ends occurs. The extent of such fragmentation varies between species, and in some species is known to be regulated by conserved chromosome breakage sequences close to the breakage sites.

Input data

• PacBio HiFi or CCS read library mapping onto MAC or MIC assembly, sorted/indexed BAM format; mapper should report valid CIGAR string and query sequences, and should perform soft clips instead of hard clips.
• Known telomeric repeat unit sequence of the genome of interest, preferably in 5’ orientation.

Identification of telomere sequences from soft-clipped sequences

When a telomere-bearing sequence of MAC origin is mapped onto a MIC reference sequence, the telomeric part does not align and is usually soft-clipped. Soft clipping operations are limited to the ends of a read, i.e. there can be no more than two soft clips on a single read.

MILTEL considers each mapped read with soft clips, and extracts the clipped segment of the query sequence, the coordinates of the clip with respect to the reference, and whether the unmapped sequence is to the left (5’) or right (3’) of the reference coordinate.

Each clipped segment extracted above is searched for the user-provided telomeric repeat sequence using NCRF, which can find tandem repeats in the presence of noise from sequencing error. Where a telomeric repeat (above a minimum length) is found, the gap distance from the beginning/end of the telomeric repeat to the clipping junction is also counted (in bp), as well as whether the telomere sequence is reverse complemented.

Output

MILTEL produces the following output files, where {OUT} is the output prefix supplied to the --out option:

• {OUT}.miltel.telomeric.gff3 - Alternative telomere addition sites / chromosome breakage sites
• {OUT}.miltel.othercalls.gff3 - Other soft- or hard-clips that do not contain telomeric sequences
• {OUT}.miltel.othercalls.fasta - Sequences of the above other clips

The GFF3 file for the alternative telomere addition sites has MILTEL in the source column (column 2). Coordinates where clipped sequence segments contain telomeres are called putative chromosome_breakage_site in the type column (column 3). Because this is a feature of zero length, the start and end fields (columns 4 and 5) are equal, and the junction is to the right of the coordinate, following GFF convention.

The score (column 6) reports the breakage score, which is the number of telomere-bearing reads clipped at that specific coordinate, divided by the total read coverage at that coordinate (as reported in the average_coverage attribute, described below).

The attributes (column 9) contain the following fields:

• ID - Unique identifier for each chromosome breakage site called.
• orientation - Whether the clipped segment is to the left or to the right of the junction.
• telomere_sense - Consensus (majority rule) on whether the clipped telomere sequences are sense (+) or reverse complement (-) of the supplied repeat unit sequence. If the supplied repeat sequence is the 5’ telomere sequence then the left orientation should correspond to the + telomere sense, and right to -.
• telomere_gap_average - Mean gap distance between beginning of telomeric repeats and the clipping junction.
• telomere_senses - String reporting the telomere sense orientations when there is more than one read clipped at that coordinate. For example, if there are 2 reads with + and 1 read with -, then the string is telomere_senses=2*+ 1*-.
• telomere_gaps - String reporting the gap distances for each clipped read. For example if there were 2 reads with gap 0 and one with gap 4, then the string is telomere_gaps=0*2 4*1
• average_coverage - Total read coverage at the coordinate, excluding secondary and supplementary mappings. Calculated the same way as the corresponding field in MILRAA output.

With the --dump option, internal data are dumped in JSON format for troubleshooting to: {OUT}.miltel.dump.json.