Skip to content

4. Extracting reads

teloclip extract writes the overhang reads to one FASTA file per contig end, so you can look at them before committing to an extension.

This step is optional but strongly recommended. It is the only stage where you see the actual sequence teloclip proposes to add.

Basic run

samtools view -h overhangs.bam \
  | teloclip extract \
      --ref-idx ref.fa.fai \
      --extract-dir split_by_contig

This writes files named <contig>_L.fasta and <contig>_R.fasta for the 5' and 3' ends respectively. The output directory is created if it does not exist.

You can pipe straight from filter without an intermediate file:

samtools view -h raw.bam \
  | teloclip filter --ref-idx ref.fa.fai --motifs TTAGGG \
  | teloclip extract --ref-idx ref.fa.fai --extract-dir split_by_contig

Richer output

samtools view -h overhangs.bam \
  | teloclip extract \
      --ref-idx ref.fa.fai \
      --extract-dir split_by_contig \
      --include-stats \
      --count-motifs TTAGGG \
      --fuzzy-count \
      --report-stats
  • --include-stats puts MAPQ, clip length and motif counts in each FASTA header
  • --count-motifs counts the given motifs in the clipped region only
  • --report-stats prints a summary, including contigs with no overhangs at all

By default the clipped portion of each read is lowercased, so the boundary between aligned and overhanging sequence is visible at a glance. Pass --no-mask to disable that.

What to look for

Open a couple of the files and read them.

Do the clips agree? Reads at the same contig end should tell the same story. Several reads whose clipped sequences align to each other is good evidence. Reads that disagree wildly suggest non-specific alignment — see Subtelomeric repeats.

Is the repeat where it should be? For the canonical vertebrate repeat you expect CCCTAA accumulating at the 5' end and TTAGGG at the 3' end. Finding the same motif at both ends in the same orientation is a warning sign.

How long are the clips? A handful of bases is not a telomere. Hundreds of bases of clean repeat is.

Is one read carrying the result? If a single read has a much longer clip than the rest, check it is not chimeric before letting extend choose it.

Contigs with no overhangs

--report-stats lists contigs where nothing was found. That is informative: a chromosome-level assembly should have overhang support at most chromosome ends. Ends with nothing may already be complete, or may be where coverage ran out.

Next: Extending contigs.