How to optimize your sequencing data for BugSeq analysis¶
The quality of the sequencing data generated in the laboratory has a direct impact on the accuracy and reliability of downstream bioinformatic results. Poor-quality or biased sequence data can reduce accuracy of de novo assembly, variant calling and estimates of microbial abundance, directly impacting organism identification, pathogen detection, outbreak analysis, and prediction of antimicrobial resistance.
Optimizing data quality through the wet-lab workflow from sample collection, extraction, library preparation and sequencing is critical to ensure that BugSeq’s bioinformatic platform delivers results you can rely on.
Collection sample quality & Storage¶
Sample quality can dramatically impact the integrity of the nucleic acids that are extracted from clinical samples. Ensuring that your samples are collected and stored appropriately for the sample type, including during transport and in the laboratory prior to extraction, can maximize the chances to recover intact nucleic acids. Ideally, this will also include minimizing the time between the collection and your processing. Transporting specimens in the appropriate media will also improve nucleic acid extraction quality.
For example, for respiratory specimens it is generally good practice to keep the samples between 2-8°C during transport. If long-term storage is required, keep most samples below -20°C but -70°C is preferable. Depending on the sample type, this sample storage may vary. Clinical and public health guidelines exist to support optimal sample storage and transport for many sample types; For example, a guideline for additional details on many common specimen types can be found here: APHL Guideline.
Repeated freeze thaws have been found to impact sequencing results including the ability to detect microorganisms from metagenomic next-generation sequencing (mNGS) data. Research sample types in advance and communicate with the sending laboratory to ensure appropriate transport of specimens, including repeated handling.
Nucleic acid extraction¶
Optimal nucleic acid extraction technique can differ depending on your targeted organism or specimen type. Use an extraction method appropriate for your target organism(s). If you are performing an agnostic sequencing method like mNGS, there should be careful consideration of the extraction method to ensure recovery of all organisms that you are targeting for detection.
Use validated platforms for pathogen detection in clinical settings and consider silica columns, mechanical lysis, or enzyme lysis in cases where recovery of all potential organisms is a priority. Careful selection of controls can also help to ensure your extraction method is recovering the desired organisms from your specimens (see the “Controls” section below for more details).
Host depletion¶
Depending on your specimen type, host nucleic acid abundance may hinder bioinformatic sensitivity. You can increase your fraction of pathogen reads by host depletion, that is removing host DNA. This can be done in several ways, and depending on your input data this may vary.
Probe hybridization depletion, human DNA depletion from methylation, rRNA depletion, or microbial enrichment may be options for host depletion. Depletion of host significantly improves BugSeq’s ability to detect important microbial genomic signatures such as antimicrobial resistance (AMR), as this is impacted by the number of microbial reads recovered in your specimen.
Library preparation¶
We recommend using a library preparation method that is validated for use within your workflow, or has been validated by the commercial vendor. Poor quality library preparation can result in low yields that reduce sequencing output, unequal representation of multiplexed samples and sequencing bias, or over-representation of host reads. The impacts of library preparation on the quality of the sequencing data and sequencing artifacts can impact BugSeq’s ability to provide informative bioinformatic results. Sequencing manufacturers often maintain detailed lists of validated library preparation workflows and protocol, including the following examples: DNA Prep (Illumina) Reference Guide or ONT Ligation Sequencing Kit Guide.
Every library preparation should involve optimization for an appropriate number of reads and depth depending on your target use case. For example, BugSeq recommends at least 40x genome coverage for whole genome sequencing of bacterial and fungal genomes. It is important to have balanced reads for your multiplexed samples to ensure a sufficient number of reads for the analysis of each specimen. The finalized library pool should be assessed before sequencing, to ensure the appropriate quantification and quality of the library will be loaded onto the sequencing instrument.
Controls¶
BugSeq recommends running appropriate controls depending on your sequencing workflow. These may include: positive, negative, or internal process controls. For negative control submission in mNGS data, you can select your negative control sample in the metadata field during BugSeq submission and BugSeq will automatically normalize data to your negative control. Running routine positive controls ensures that the extraction, library preparation, and sequencing was not a failure.
Using these controls can enable you to leverage BugSeq’s analytics to accurately identify where the failure point in analysis could be, and can also help laboratories identify wet-lab failure points.
A unidirectional molecular workflow is recommended for NGS assays to ensure accurate results and prevent the contamination of sequencing samples. From reagent preparation, sample processing, extraction, amplification to library preparation it is imperative to maintain correct laboratory practices for contamination prevention. This includes proper forward to backward movement and separation of clean and amplicon contaminated areas, proper PPE (personal and equipment), and use of effective decontamination methods for nucleic acids.
Instrumentation¶
Using a serviced sequencing instrument (Illumina (including short-read sequencers), Oxford Nanopore Technologies, Ion Torrent and PacBio) will improve the ability to recover accurate insights from BugSeq’s latest software, as well as using updated flowcell and sequencing chemistries. High quality sequencing data starts with a well-maintained instrument and up-to-date chemistry. These upstream errors have a direct impact on bioinformatic analysis and can cause errors in your pathogen detection, AMR detection, pathogen abundance and more.
For optimal BugSeq upload performance of ONT sequencing, we recommend to use fewer and larger FASTQ files (>100,000 reads rather than the default of 4,000). For Illumina sequencing - assuming paired end barcoding - we recommend keeping the paired-end files separate, as BugSeq will automatically merge these files during the first stages of data analysis. Use barcode aliases to assign custom sample names, which BugSeq’s software will automatically recognize. Alternatively, submit one FASTQ file per sample, with the sample name derived from the filename. Multiple barcode FASTQs can also be combined into a single sample by concatenating the files before submission.
Selecting the Appropriate Analysis Configuration Settings¶
Setting default submission parameters in the BugSeq “Labs” setting allows you to specify the specimen type, sequencing platform, database options for default submissions, as well as quality control settings for analyses, who is allowed to submit and review analyses, and more.
It is recommended to submit data directly through BugSeq’s integration with ONT’s EPI2ME or Illumina’s BaseSpace for faster and more automated data processing. It is also recommended that you submit raw, unfiltered, sequencing data, as BugSeq’s pipeline automatically performs quality controls, trimming and filtering. Specify the correct sample type, sequencer details, and sequencing material to optimize BugSeq’s analysis and ensure you receive timely and correct results.
BugSeq offers multiple pipelines (mNGS, isolate based WGS, 16S, ITS and viral amplicon sequencing), and different sample preparation workflows require different analyses. Options chosen in the submission will affect the tools used for assembly and classification with BugSeq’s software. For example, different classification workflows perform differently for RNA versus DNA and could lead to flagging your desired organisms as contaminants or a mismatch between your sample type and expected results.
Reach out¶
Beyond our bioinformatics pipelines, BugSeq offers consulting services to support your workflow from sample preparation through to sequencing. We help optimize each step of the process so you can generate high-quality data and get the most reliable results from your BugSeq analysis. Reach out to contact@bugseq.com for more information on our consulting services.