Introduction
Welcome back! We’re excited for you to read the latest edition of ‘From the Computer to the Clinic’
In this newsletter, we are exploring how research in bioinformatics and computational biology can drive clinical progress. By sharing success stories in one disease area or domain of research, we aim to inspire the use of these successful approaches for other diseases and research areas also.
If you haven’t already, you can subscribe to this newsletter, or share it with friends and colleagues
Featured Study
Drugs of abuse – like fentanyl and other synthetic opioids – take over 100,000 lives each year in the US, and many additional lives abroadA. Drug overdoses require timely and effective treatment. This would be difficult enough if it were easy to diagnose and treat an overdose once a victim is identified. Unfortunately, it is not. Criminal chemists in clandestine laboratories are constantly synthesizing new variants of well-known drug compounds, making diagnosis more difficult and complicating treatment with new and dangerous effects. To meet the public health crisis of drug abuse, researchers are collaborating with both the medical community and law enforcement to monitor the spread of drugs and treat overdose victims more effectively.
One recent study by a team of researchers, doctors, and forensic scientists from British Columbia has developed a novel computational approach to help local clinics detect what drugs are proliferating in their area. This information will help clinics marshal the necessary resources to diagnose and treat overdose victimsB. It will also help law enforcement get a better hold on what new kinds of drugs are being produced and where they are being produced.
When a clinic is trying to determine which drug a patient has in their system, they often use a machine called a mass spectrometer. This device can detect chemical signatures that identify individual compounds – like drugs of abuse – amidst the complex mixture of chemicals present in urine or other bodily fluids. To be certain about what drug is present however, clinics acquire reference standards – pure solutions of drugs – and run them on the same mass spectrometer so that they know what chemical signatures they should be looking for.
The issue is that there are a lot of different drugs of abuse, and new variants frequently emerge. Acquiring reference standards for all of these drugs would be prohibitively expensive. Often, a tough decision must be made about what references are going to be acquired. This decision is often made using the national prevalence of drugs of abuse as input data. This approach is prone to missing local trends in drug prevalence, and a clinic may fail to acquire reference standards for drugs that are relevant to their area.
An attractive alternative would be for a clinic to re-analyze its own data records – the mass spectrometry results on file for past patients – and find out what drugs are most prevalent in their clinical samples. The authors of this edition’s featured study took this idea and came up with a data-driven way for clinics to make decisions about which reference standards to purchase. Clinics that use this approach can maximize their preparedness to treat overdose victims.
Featured Study: Identification of Emerging Novel Psychoactive Substances by Retrospective Analysis of Population-Scale Mass Spectrometry Data Sets (Skinnider et al., Analytical Chemistry, 2023)
To accomplish this project, the researchers first needed to build a database of reference standards for drugs of abuse. They gathered information on 83 drugs of abuse that have recently been seized in Canada. For each of these drugs, the researchers had a chemical signature – the mass spectrum.
A mass spectrum is like a barcode for molecules, containing a set of peaks arranged from left to right (these peaks correspond to fragments of a drug compound with different weights). Each of these peaks also has a certain height (referred to as its intensity). Thus, the mass spectra of two molecules may differ both in the location of the peaks from left to right, and the height of their peaks – and this is how you can tell them apartC.
With the database of drug spectra in hand, the researchers screened over 12,000 urine samples from the Provincial Toxicology Centre at the British Columbia Centre for Disease Control. These samples were collected over three years from August 2019 to August 2022. Each sample had associated mass spectrometry data that could be screened for matches to drugs in the 83-drug database.
When a donor sample mass spectrum had at least two peaks in common with the mass spectrum of a drug in the donor database, the researchers reasoned that the drug could be present in the sample. The researchers also performed further analysis to evaluate the confidence of each matchD.
Based on these criteria, some drugs appear to be very common in the clinical samples. Ten of the 83 drugs were found in 20 or more patient samples – including the opioid fluorofentanyl, and bromazolam, which lies in the same class of drugs as Valium and Xanax. The researchers were also able to visualize trends in the prevalence of these drugs over time. Flourofentanyl, for example, was found in about 12% of samples at the end of the study period (2022), whereas it was nonexistent at the start of the study period in 2019.
Though the drug detection approach outlined in the paper can be optimized – for example, by developing more precise methods for detecting drugs in forensic samples – the results are very promising.
Based on this approach, the authors envision a regular surveillance system that monitors the emergence of new drugs in a particular area, perhaps on a monthly basis. This system could be implemented at other local locations beyond the BC Provincial Toxicology Centre that have clinical samples to analyze. To facilitate broader implementation, the authors have created open source software packages that local centers could use to perform the analysis shown in the paper on their own.
One key aspect of this system is that local centers can use the database of 83 drug reference standards in the study as a global compendium for drugs of abuse. They don’t have to develop similar databases themselves – just screen their samples to see what drugs are present, and purchase a relatively small number of reference standards for the most relevant local drugs. For this system to work over the long term, the global compendium will need to be expanded as new drugs are seized – otherwise they won’t be detected in the samples of local clinics. How this effort could be organized (i.e., whether the responsibility for expanding the database falls solely on a central organization or whether it is a collaborative effort between local provinces or states) is an open question.
A network of local centers monitoring emerging drugs of abuse would produce useful data about where certain kinds of drugs are emerging and even how new drugs typically spread between regions. This information is useful not just for preparing medical professionals and assembling medical supplies to treat a local population of drug users – but also for helping law enforcement pin down the sources and prevent the spread of illicit drugs.
Broader Trends
(A) For readers in the US (most of our current audience), according to the National Institute on Drug Abuse, over 106,000 people died from a drug overdose in the US alone during 2021. The death totals have risen dramatically since the early 2000s. More recent estimates suggest that the death total was over 109,000 from May 2021-May 2022, and over 112,000 from May 2022-May 2023. The numbers continue to rise. Currently, synthetic opioids take the most lives by far of any drug class in the US.
(B) Scientific research can be very impersonal. To get a true feel for the magnitude of the opioid epidemic, it is worth reading the more personal accounts of clinics who are treating overdose victims (this story is from an emergency department in Philadelphia) – or the experience of overdose victims themselves (this website documents the experiences of overdose victims in their own words)
(C) Here is a relatively accessible review article on how mass spectrometry works in general. This video also gives a good overview on tandem mass spectrometry, the technique used in the featured paper.
(D) Beyond the scope of this article, but worth reading more about in the results section of the paper (see the section of the results titled “Refinement of Tentative NPS Identifications by Mass Spectral and Chromatographic Data”)