TELOMORE

Telomore is a tool for identifying and extracting telomeric sequences from Oxford Nanopore or Illumina sequencing reads of Streptomycetes spp. that have been excluded from a de novo assembly. It processes sequencing data to extend assemblies, generate quality control (QC) maps, and produce finalized assemblies with the telomere/recessed bases included.

Table of Contents

• Installation
• Before running Telomore
• Usage
• Process Overview
• Outputs
• Evaluating Output and Troubleshooting
  • Examples
  • Troubleshooting

Installation

The easiest way to install telomore is using pip:

pip install telomore

Telomore depends on a number of CLI tools:

• Bowtie2
• Emboss tools (cons specifically)
• Lamassemble
• LAST-DB
• Mafft
• Minimap2, version 2.25 or higher
• Samtools

These can be installed using the conda recipe in this repo:

conda env create -f environment.yml -y

This repo can also be downloaded and used to install telomore:

# Activate telomore conda env
conda activate telomore

# Clone telomore repo
git clone https://github.com/dalofa/telomore && cd telomore

# Install package
pip install -e '.[dev]'

Before running Telomore

Telomore does not identify linear contigs but rather rely on the user to provide that information in the header of the fasta-reference file.

To capture the archetypal (or other actinobacterial) telomere in full in an otherwise complete assembly, it is often sufficient to extend the assembly using Illumina data. However, in certain cases large chunks of an arm is missing and thus, one must extend with Nanopore before extending with Illumina data.

Beware: Telomore was designed with expectation of a complete/near-complete assembly, where each replicon is represented by a single contig. While, a more fragmented assembly can be input, this might lead to the overrepresentation of sequence.

Usage

telomore --mode <mode> --reference <reference.fasta> [options]

Required Arguments

• --mode Specify the sequencing platform. Options: nanopore or illumina.
• --reference Path to the reference genome file in FASTA format.

Nanopore-Specific Arguments

• --single Path to a single gzipped FASTQ file containing Nanopore reads.

Illumina-Specific Arguments

• --read1 Path to gzipped FASTQ file for Illumina read 1.
• --read2 Path to gzipped FASTQ file for Illumina read 2.

Optional Arguments

• --coverage_threshold Set the threshold for coverage to stop trimming during consensus trimming (Default is coverage=5 for ONT reads and coverage=1 for Illumina reads).
• --quality_threshold Set the Q-score required to count a read position in the coverage calculation during consensus trimming (Default is Q-score=10 for ONT reads and Q-score=30 for Illumina reads).
• --threads Number of threads to use (default: 1).
• --keep Retain intermediate files (default: False).
• --quiet Suppress console logging.
• --skip_side Skip processing of a terminal side. Accepts left or right.

Example (skip left side):

telomore --mode nanopore --single reads.fastq.gz --reference genome.fasta --skip_side left

Process overview

The process is as follows:

1. Map Reads: Reads are mapped against all contigs in a reference using either minimap2 or Bowtie2.
2. Extract Extending Reads Extending reads that are mapped to the ends of linear contigs are extracted.
3. Build Consensus The terminal extending reads from each end is used to construct a consensus using either lamassemble or mafft + EMBOSS cons
4. Align and Attach consensus The consensus for each end is aligned to the reference and used to extend it.
5. Trim Extended Replicon In a final step, all terminally mapped reads are mapped to the new extended reference and used to trim away spurious sequence, based on read-support.

Outputs

At the end of a run Telomore produces the following outputs:

├── {fasta_basename}_{seqtype}_telomore
│   ├── {contig_name}_telomore_extended.fasta
│   ├── {contig_name}_telomore_ext_{seqtype}.log
│   ├── {contig_name}_telomore_QC.bam
│   ├── {contig_name}_telomore_QC.bam.bai
│   ├── {contig_name}_telomore_untrimmed.fasta
│   └── {fasta_basename}_telomore.fasta
└── telomore.log # log containing run information.

In the folder there is a number of files generated for each contig considered:

File Name	Description
{contig_name}_telomore_extended.fasta	Original contig sequence + added terminal bases - trimmed bases
{contig_name}_telomore_ext_{seqtype}.log	Log contianing information about bases added, trimmed off and final result.
{contig_name}_telomore_QC.bam	BAM file containing terminal reads mapped to {contig_name}_telomore_extended.fasta. Useful for manual inspection of the extension
{contig_name}_telomore_QC.bam.bai	Index file for the corresponding BAM file.
{contig_name}_telomore_untrimmed.fasta	Original contig sequence + added terminal bases

Additionally, there is a fasta-file collecting all tagged linear contigs as they appear in {contig_name}_telomore_extended.fasta together with all non-linear contigs in the order they appear in the original file.

Inspecting the {contig_name}_QC.bam-file in IGV (Integrative Genomics Viewer) can be informative in evaluating the extended contig.

Evaluating Output and Troubleshooting

The best way to evaluate whether your assembly have been extending in a meaningful way is to inspect the QC-bam file produced at the end of the Telomore workflow. Ideally you would a high degree of agreement between the consensus and the reads used to build it. Below are some examples:

Examples

Example 1: Illumina extension of NBC_00015

The chromosome of Streptomyces sp. NBC_00015 were extended with ONT data and subsequently with Illumina data. The Illumina data extended the left end with 23 bases, which look well supported when inspecting the bam file at the end.

Example 2: Nanopore extension of NBC_00008

The chromsome of Streptomyces sp. NBC_00008 were extended with ONT data. The right end were extended with 23 bases, which seem well supported by the aligned reads. However, about 7 bases were trimmed of the consensus due to low quality of the final bases of the reads. Thus, one could consider other sequencing methods or rerunning with a decreased threshold for trimming.

Example 3: Hairpin-type telomeres (closed telomeres)

Certain bacteria (Such as Borrelia) utilize a type of telomere with covalently closed hairpins, rather than a blunt-end with a protein. These telomeres have the opposite problem, where a chimera is formed by readthrough of the hairpin onto the complementary strand. Using such reads will overextend the genome with complementary strand since the reads themselves support it:

See #51 for an example (which is also where this example originates from). [Autocycler](https://github.com/rrwick/Autocycler/wiki/Linear-sequences) provides options for dealing with this issue. For mixed cases, where a hairpin-type telomere is on one end and streptomyces type telomere is on the end skip-side can be used to control which side are extended.

Troubleshooting

In cases where a consensus is rejected it can be worthwhile to investigate the particular reason. Typically, the reasons are as follows:

• There are no reads to construct a consensus from.

No reads extend the contig at the side.

• {side} consensus does not extend genome This can be either due to a consensus which maps to the genome in a non-extending fashion or a consensus which does not map to the genome. A non-mapping consensus is often due to spurious or chimeric reads. By running telomore with the --keep option, the reads used to construct each consensus can be found in the respective .aln-files. Additionally, the original mapping file {reference}_map.bam can be inspected to see if any of the reads seem out of place. For example:

In this case one read - which seems to be chimeric based on BLAST-matches - will yield a consensus that does not map at the end. Similarly, reads of poor quality present at the end can generate consensus that does not map at the or does not extend. Removing these reads can fix the issue but should be done with care.