🦠 microbiology

SELECT 2.0: Refined and open access SELection Endpoints in Communities of bacTeria (SELECT) method to determine concentrations of antibiotics that may select for antimicrobial resistance in the environment

This paper introduces the refined, open-access SELECT 2.0 method to empirically determine predicted no-effect concentrations for resistance (PNECRs) for 32 antibiotics, enabling standardized environmental risk assessments to rank antibiotics by their potential to select for antimicrobial resistance in various ecosystems.

Original authors: Hayes, A., Kay, S., Lowe, C., Gaze, W. H., Recker, M., Buckling, A., Murray, A. K.

Published 2026-03-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Hayes, A., Kay, S., Lowe, C., Gaze, W. H., Recker, M., Buckling, A., Murray, A. K.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine a world where the tiny, invisible bacteria living in our rivers and sewage systems are constantly training for a war. They aren't training to fight us, but to survive the "weapons" we use against them: antibiotics. When we flush leftover medicine down the toilet or when farms wash manure into rivers, we leave behind tiny traces of these drugs. Even at very low levels, these traces act like a gym for bacteria, teaching them how to become super-strong and resistant.

This paper is about a new, smarter way to measure exactly how much "gym time" (antibiotic concentration) it takes to turn a normal bacterium into a super-resistant one.

Here is the breakdown of the study using simple analogies:

1. The Problem: The "Silent Gym"

For years, scientists have worried that the environment is full of low-level antibiotics. The big question is: How much antibiotic is enough to start training bacteria to resist us?

Old Way: Imagine trying to find the exact weight that makes a person lift a barbell. The old method (SELECT 1.0) was like asking, "Did they lift it? Yes or No?" It only checked specific weights (like 10kg, 20kg, 30kg). If the magic weight was 15kg, the old method might miss it entirely or guess wrong because it only looked at the big jumps.
The Risk: If we guess the "safe" level too high, we might let dangerous amounts of antibiotics into the environment, unknowingly training bacteria to become immune to our medicine.

2. The Solution: SELECT 2.0 (The "Smart Coach")

The researchers created an upgraded version of their test called SELECT 2.0. Think of this as upgrading from a simple "Yes/No" checklist to a high-tech smart coach that watches every single movement.

The Experiment: Instead of using just one type of bacteria (like a single athlete), they used a whole crowd of different bacteria from sewage. This is like testing a whole sports team instead of just one player, because in the real world, bacteria live in messy, complex communities.
The New Math: The old method looked for a sudden drop in growth. The new method (SELECT 2.0) draws a smooth curve through the data. It asks: "At what tiny, almost invisible amount does the bacteria's growth slow down by just 1%?"
Why 1%? Because bacteria multiply incredibly fast. Even a tiny 1% slowdown in a population of millions is a huge deal. It's like noticing a single drop of water in a swimming pool; if you ignore it, the pool eventually overflows. This makes the new test stricter and safer for the environment.

3. The Results: Who is the Biggest Threat?

The team tested 32 different antibiotics (the "weapons") to see which ones were the most dangerous at low doses.

The "Super-Strong" Trainers: They found that Ciprofloxacin (a common antibiotic) and Ceftriaxone are the most dangerous. They can start training bacteria to resist them at concentrations so low they are almost impossible to measure with old tools. It's like a whisper that is loud enough to wake up a sleeping giant.
The "Weak" Trainers: Antibiotics like Vancomycin and Penicillin needed much higher doses to start training the bacteria. They are like a gentle breeze that doesn't move the giant.
The Big Database: This study created the largest list of these "danger zones" ever made using a single, consistent method. It's like finally having a complete map of all the potholes on a road, rather than just guessing where they might be.

4. The Real-World Check: Is Our Water Safe?

The researchers took their new "danger map" and compared it to real water samples from the UK and around the world.

The Findings: In many places, the water contains enough antibiotics to be in the "danger zone."
The Villain: Ciprofloxacin was the biggest offender. In both the water going into sewage plants and the water coming out, the levels were high enough to train bacteria to resist it.
The Good News: Not every antibiotic is a problem. Some are safe at current environmental levels. This helps regulators know exactly which drugs need stricter controls.

5. Why This Matters (The "So What?")

Imagine you are a city planner trying to keep a bridge safe.

Before: You guessed where the cracks were. You might have missed a small crack that eventually caused the bridge to collapse.
Now: With SELECT 2.0, you have a laser scanner that finds the tiniest hairline fracture. You can fix it before it becomes a disaster.

The Bottom Line:
This paper gives us a better, faster, and cheaper ruler to measure the danger of antibiotics in our environment. It proves that we need to be much more careful with certain drugs (like Ciprofloxacin) because even tiny amounts in our water are enough to create "super-bacteria" that could make our medicines useless. By using this new method, we can set better rules to protect our water and, ultimately, our health.

The Takeaway for Everyone:
We can't just flush our medicine away and hope for the best. This study shows us exactly how much is too much, so we can stop training the bacteria to win the war against us.

1. Problem Statement

Antimicrobial resistance (AMR) is a critical global health threat, with environmental pollution acting as a significant driver for the evolution and dissemination of resistant bacteria. Current Environmental Risk Assessments (ERAs) for antibiotics rely heavily on ecotoxicological endpoints (e.g., effects on algae or fish) which often fail to detect the low concentrations of antibiotics that select for AMR.

The Gap: Existing methods to determine the Minimal Selective Concentration (MSC)—the lowest concentration of an antibiotic that selects for resistance—are often limited to single-species competition experiments or specific resistance markers. These approaches suffer from heterogeneity, making cross-study comparisons difficult.
The Need: There is a lack of standardized, high-throughput, and cost-effective methods to generate Predicted No Effect Concentrations for Resistance (PNECRs) across a wide range of antibiotics using complex bacterial communities that better mimic environmental conditions.

2. Methodology: SELECT 2.0

The authors refined a previously published method called SELECT (SELection Endpoints in Communities of bacTeria) into SELECT 2.0.

Experimental Design:
- Inoculum: Untreated sewage samples (representing a complex bacterial community) were collected from a wastewater treatment plant, cryopreserved, and resuspended in sterile Iso-sensitest broth.
- Assay: The community was exposed to a range of antibiotic concentrations (up to 32 different antibiotics spanning 11 classes) in 96-well plates.
- Measurement: Optical Density at 600nm (OD600) was measured every 10 minutes for 12 hours using a plate reader. This high-frequency kinetic reading replaces the hourly measurements of the previous version to generate smoother growth curves.
- Controls: Included no-antibiotic controls, solvent controls (specifically for NaOH), and sterile broth controls.
Statistical Refinement (The Core Innovation):
- SELECT 1.0: Relied on binary statistical significance (Dunn's test) to identify a Lowest Observed Effect Concentration (LOEC). This was sensitive to random error and the specific spacing of test concentrations.
- SELECT 2.0: Fits a four-parameter log-logistic dose-response curve to the growth data. It calculates the EC1 (Effective Concentration causing a 1% reduction in maximum growth density).
- Rationale for EC1: The authors argue that EC1 is more protective than the standard EC10 used in ecotoxicology because bacteria grow rapidly and in vast numbers; even a 1% shift in a population of millions represents a significant increase in resistant bacteria.
- PNECR Calculation: The PNECR is derived by dividing the EC1 by an Assessment Factor (AF) of 10, aligning with European Medicines Agency (EMA) recommendations for aquatic risk assessment.
Validation & Risk Assessment:
- Comparison: Data was analyzed using both SELECT 1.0 and 2.0 methods and compared via Bland-Altman analysis to assess agreement.
- Cross-Study Comparison: Results were compared against a meta-analysis of PNECRs from various other experimental systems.
- Environmental Risk Assessment (ERA): PNECRs were compared against Measured Environmental Concentrations (MECs) from two datasets:
  1. UKWIR CIP3: Wastewater influent/effluent from England and Wales.
  2. UBA: Global wastewater data.
- Risk Quotient (RQ): Calculated as $RQ = MEC / PNECR$. An RQ $\ge$ 1 indicates high risk.

3. Key Contributions

Methodological Refinement: Transitioned from a binary LOEC approach to a continuous, model-based EC1 approach, reducing the impact of random error and allowing for the prediction of effect concentrations outside the tested range.
Open Science: The authors released all statistical code (R), raw data, and a tutorial on GitHub/Zenodo to ensure reproducibility and facilitate adoption by other laboratories.
Comprehensive Database: Generated the largest single-method experimental database of PNECRs to date, covering 32 antibiotics across 11 classes (including beta-lactams, quinolones, macrolides, etc.).
Standardization: Provided a standardized protocol using a complex sewage community, addressing the issue of heterogeneity in previous AMR selection studies.

4. Key Results

Method Comparison: Bland-Altman analysis showed that SELECT 1.0 and 2.0 results generally agreed, with all data points falling within the limits of agreement. However, SELECT 2.0 generally produced more protective (lower) PNECRs for highly selective antibiotics (e.g., Ceftriaxone, Ciprofloxacin) by detecting subtle growth reductions that LOEC missed.
Antibiotic Selectivity:
- Most Selective (Lowest PNECRs): Quinolones (e.g., Ciprofloxacin PNECR: 0.005 µg/L) and Beta-lactams (e.g., Ceftriaxone PNECR: 0.00003 µg/L).
- Least Selective (Highest PNECRs): Glycopeptides (Vancomycin PNECR: 954 µg/L) and Penicillins.
- Variation: Significant intra-class variation was observed in Quinolones and Beta-lactams, suggesting that risk assessments cannot rely solely on class-level data.
Environmental Risk Assessment (ERA):
- Global Data: 12 antibiotics posed a high selection risk (RQ > 1) in urban wastewater. Ciprofloxacin was the most critical, showing high risk in both influent and effluent.
- UK Data: Ciprofloxacin was identified as the highest risk pollutant in English and Welsh wastewater, with RQs > 1 in both influent and effluent. Clarithromycin, Erythromycin, Sulfamethoxazole, and Trimethoprim also showed elevated risks.
- Conclusion: Antibiotic pollution in wastewater, even in high-income countries, poses a significant risk for selecting AMR.

5. Significance and Future Directions

Policy Impact: The generated PNECRs provide a robust scientific basis for setting Environmental Quality Standards (EQS) for antibiotics, moving beyond traditional ecotoxicology to directly address AMR selection risks.
Stewardship: The data helps prioritize which antibiotics require stricter discharge controls or manufacturing limits to prevent environmental selection of resistance.
Limitations & Future Work:
- Genotypic Validation: The study used phenotypic growth reduction as a proxy for genotypic selection. Future work should validate these PNECRs against long-term (week-long) selection experiments tracking specific resistance genes (e.g., intI1).
- Inoculum Diversity: The method currently uses sewage inoculum. Future studies should test river water or soil communities to determine if PNECRs vary by environment.
- Mixture Effects: The current method tests single antibiotics. Future iterations should address the "cocktail effect" of multiple antibiotics and co-selective compounds in the environment.

In summary, SELECT 2.0 represents a significant advancement in environmental AMR surveillance, offering a rapid, reproducible, and open-access tool to quantify the selective pressure of antibiotics in the environment, thereby supporting more effective One Health risk assessments.