• Filters
  • only
  • only
  • only
  • only
  • only
  • only
  • only
  • only
  • only

Select All / Deselect All

Reset

Karius Research: How to Advance Basic Science From Within Industry

Young scientists just starting out might have the perception that academic science makes all the research discoveries, while industry optimizes and sells them.

We’d like to suggest that the flow of value can go both ways.

True, it was academic research in Steve Quake’s lab at Stanford that led to the founding of Karius: the discovery that microbes leave traces of their genetic material in the blood as cell-free DNA. And academia is certainly a great place to be.

But we’re demonstrating that industry has its own significant role to play in advancing basic science and accelerating research discoveries.

Why is this?

First of all, there is often more money available in industry that can be focused on characterizing and tackling any problems that arise.  Not to mention the high robustness standards from oversight agencies that require you to rigorously test for biases and bottlenecks.

While these long, expensive experiments might be uncommon for many academic labs, they do often generate discoveries that are valuable to the broader community.  

And the stakes are higher when you know your research will directly impact the quality of care that someone is going to receive, so there’s an extra layer of motivation to make sure you understand every aspect of your product.

Of course there are barriers being in industry that limit us from freely sharing every one of the discoveries we’ve made. But we can raise awareness about the new knowledge we’re generating, talk about as much of what we’ve learned as we can, and collaborate where possible with academics and clinicians.

Here are three examples of how our work here at Karius is advancing scientific research:

Environmental Contamination

After removing human DNA from the picture, the majority of microbial DNA signals left actually derive from reagents and plastics in the environment. These contaminants may include clinically relevant microbes not present in the plasma sample that could confound the test results.

They also confound many metagenomics studies, in that unless the environmental contamination is addressed and removed, the resulting community profile can be partly or even largely derived from the DNA contaminating the reagents used.

Which is a potentially huge problem when a patient’s life is on the line.

We have put a tremendous amount of focus into reducing environmental contamination (EC) in our assay. Our clinical approach includes a four-part method of reduction and accounting: there are four EC controls on every run, our algorithms model EC expectations, the most contaminated reagents are replaced, and the irreplaceable reagents undergo a novel DNA/RNA cleaning procedure.

As a result of these modifications, an average of only <1 read per sample is attributable to EC, while the reads attributed to plasma-derived signal is typically in the hundreds or thousands.

Noise

For infectious disease detection, the pathogen signal in plasma is about one millionth the level of host-derived signal. The signal for individual pathogens is often much less than that, unless the patient is very sick. This makes it hard for widely available technologies to be able to extract meaningful information.

Karius is the first team to discover and characterize this microbial cell-free DNA signal and use it to detect pathogens.

To be able to do this, we developed a proprietary one-hour enrichment protocol that results in 100-fold increased signal-to-noise and 100-fold reduced sequencing cost compared to standard next-generation sequencing (NGS) protocols. Instead of using targeted amplification, pull-downs, or other methods that target particular sequences, we focus instead on removing human DNA and seeing everything else, hypothesis-free.

We also established a reference range based on a self-reported asymptomatic cohort to allow us to call signals as statistically relevant or not. All new test results are compared first to the background we see in reagents and the environment to determine significance, and then to this reference range to help inform clinical relevance.

Bias

One of the biggest challenges in reliably delivering a test that probes for the presence of over a thousand different microbes is to ensure similar test performance across such a diverse spectrum of microbes that differ in GC-content, super-kingdom, prevalence as commensals in the body, EC levels, and many other aspects.

In order to maintain robustness at high throughput, Karius bioinformatics pipelines calculate relative abundance of microbes from the NGS results, with extensive bias monitoring and correction features.

We control, measure, and account for bias in every sample, giving us far better sensitivity and precision than standard NGS protocols.

Some of the controls we use in the lab that assist our pipelines in this effort include sample tracking spike-ins, quality control checks for biases in each sample, reduction and characterization of the GC bias common in NGS protocols, and methods to track carryover contamination between sample neighbors on our automated processing platforms.

Looking forward

Karius can detect more than 1,000 species of pathogens today – from deep-seated and challenging infections including co-infections, unculturable ones, evolving microbes, and in the presence of antibiotics – with next day results. Innovating in our domain and distributing knowledge back to the scientific community are principles we strive to live by at Karius.

We believe we have built a fundamentally better way to approach infectious diseases that benefits patients and physicians while powering basic science.

The challenges we have overcome are significant, and we’re proud to have developed a test with clinical-grade breadth, depth, strength, and speed that is positively impacting patients’ lives.

But there’s still a long way to go.

Please email help@kariusdx.com to join us in this endeavor. We’d love to hear from you.