Most antibiotics that hit the ribosome, the protein-making machine inside every bacterial cell, attack the same few sites. The ribosome is the target of about one third of all antibiotics prescribed today. Bacteria have had decades to evolve defenses against those sites. A new study in Nature describes a compound that goes somewhere else. technologynetworks

What was found
The compound is called manikomycin. It is a cyclic depsipeptide made by Streptomyces rimosus, a soil bacterium that has long been known as the source of oxytetracycline. The authors describe it as the first antibacterial agent, to their knowledge, to target the E-site in the large ribosomal subunit. rapidmicrobiologypaperity

How it works: the assembly-line picture
Think of the ribosome as an assembly line. Carrier molecules (tRNAs) bring in building blocks, then leave through an exit. Manikomycin binds that exit site and interferes with translocation, the step in which the ribosome moves along its message. The used carrier can’t leave, the machine stalls and protein-making stops. paperityearth

Why a new target matters
Because medicine has never put selective pressure on this site, bacteria have no existing resistance mechanisms aimed at it, according to the lead researcher. The drug also enters bacterial cells through several transport routes, which may make resistance harder to develop. mcmastertechnologynetworks

What it was active against
Early results show activity against priority pathogens including Salmonella, E. coli and Klebsiella. mcmaster

Why it spares human cells
Human ribosomes carry an extra protein right where manikomycin would need to bind, which appears to keep the drug out. Even at high doses, the drug didn’t kill the human cells tested or damage blood cells. earth

The catch: it isn’t a medicine yet
The compound doesn’t stay in the bloodstream long enough to kill bacteria efficiently in animals or humans, so several things need to improve first. The team is now working to move it toward clinical development. Many candidates fail at this stage, so it is best seen as a promising lead, not a coming cure. technologynetworksmcmaster

The bigger picture
This is the fourth new antibiotic candidate from the same lab in just over a year. You can read about one of the others in our post on lariocidin. Together they suggest that natural sources still have a lot to offer. mcmaster

What it means for laboratories
If a drug like this ever reaches clinics, laboratories will need validated susceptibility testing methods and breakpoints for it. As we covered in our verification vs validation post, new methods must be shown to work in your own lab before results are reported.

Reference
Kaur M, Travin DY, Coombes (and colleagues), Wright GD. A natural depsipeptide antibiotic binds the E-site of the bacterial ribosome. Nature, 2026. DOI: 10.1038/s41586-026-10589-2. https://www.nature.com/articles/s41586-026-10589-2

Related reading: lariocidin-new antibiotic candidate and verification vs validation .
Next post (Friday): measurement uncertainty in microbiology, explained simply.

In our last post we looked at a CRISPR test that can identify bacteria and resistance genes from a positive blood culture. The BADLOCK study reported a one-pot test covering nine bacterial species and four major resistance genes, run with only a heat block and a basic fluorometer. Impressive, but here is the question every laboratory manager should ask next: how do we know it will work in our laboratory? broadinstitute

The answer is verification and validation. The two words sound alike, but they answer different questions.

In simple words

• Verification: “Does this approved method perform as claimed in our hands?”
• Validation: “Is this method fit for the purpose we intend?” It is needed when a method is new, developed in-house, or changed from its original use.

A good rule: if you follow the manufacturer’s instructions exactly, you verify. If you change anything, such as the sample type, the organisms covered or the procedure, you validate.

The same test can need either
Suppose a lab wants to use a BADLOCK-style assay. The published work used positive blood cultures. If the lab used it on those samples as described, verification may be enough. If the lab wanted to run it on urine or direct blood, that is a new purpose, so it needs validation.

Step 1: Define the purpose first
Write down the sample type, which organisms and genes you will report, and the performance you need. This matters because a test result must match the question being asked. For example, detecting a resistance gene does not always mean the organism is resistant in practice, so the report should be worded carefully, and susceptibility testing may still be needed.

Step 2: Choose what to measure
For a qualitative (yes/no) test, labs usually check:

• Agreement with a reference method, using known positives and negatives
• Limit of detection (LoD): the lowest amount of target the test reliably finds. It is commonly defined as the level detected in at least 95% of repeat tests, for example 19 of 20
• Specificity: does it avoid false positives with similar organisms?
• Repeatability: same result on repeat runs and different days or operators

Step 3: A worked example (invented numbers, for teaching)
Say you test 40 samples known to be positive and 20 known to be negative, using a reference method as the “truth”.

• New test finds 38 of 40 positives: positive percent agreement = 38/40 = 95%
• New test is negative in 20 of 20 negatives: negative percent agreement = 20/20 = 100%
• Overall agreement = 58/60 = 96.7%

Before you start, your acceptance criterion might have been “at least 90% agreement.” You pass, but look closer.

1. The two missed positives matter. Investigate them. Were they low-level samples or one organism type? This is how you discover the test’s weak spot.
2. Small numbers mean wide uncertainty. A true 95% measured on only 40 samples could realistically be anywhere from about 83% to 99%. That is why more samples give more confidence.

Step 4: Fix criteria before testing
Imagine the acceptance limit had been “95% or better” and you saw 89.5%. It would be tempting to lower the bar. In the BADLOCK paper, 274 of 306 assays (89.5%) fully or partly agreed with the gold-standard results. Whether that is good enough depends on the purpose and the risk. That decision belongs before the experiment, written in your plan, and not afterwards. broadinstitute

Step 5: Record and approve
Keep raw data, lot numbers, operator names and dates. The authorized person signs the report, the method enters document control, and staff are trained and shown to be competent before reporting patient results.

When to repeat it
Re-verify after major changes, such as new equipment, a new reagent lot type, a new sample type or a software update.

Where to look for guidance
Many labs follow CLSI guidance for qualitative tests and detection limits, and ISO 16140 for food microbiology methods, alongside ISO 15189:2022 and ISO 17025:2017. In ISO 15189:2022 (medical laboratories), verification is covered by clause 7.3.2 and validation by clause 7.3.3. In ISO/IEC 17025:2017 (testing and calibration laboratories), verification is covered by clause 7.2.1.5 and validation by clause 7.2.2. Medical and clinical laboratories follow ISO 15189, while food, environmental and other testing laboratories follow ISO 17025.

Quick check

1. A lab wants to run a blood-culture assay on urine. Verification or validation?
2. Why does a 95% result from only 40 samples not mean the test is truly 95%?
3. Why must acceptance criteria be written before the study?

Related reading: The Myth of Document Control | The Illusion of Control: Risk Management | CRISPR diagnostics .
Next post (Tuesday): a new antibiotic that targets the ribosome in a new way.

Identifying the bacteria in a patient’s blood, and which antibiotics they resist, still takes days in many laboratories. That delay matters most in regions with limited diagnostic equipment.

A new study in Science Advances (October 2026) describes a possible shortcut. The test, called BADLOCK, is a one-pot CRISPR-Cas13a reaction that detects nine bacterial species and four major resistance genes directly from positive blood culture, using only a heat block and a basic fluorometer. broadinstitute

How it works
CRISPR was first found as a bacterial defense system. Here, the Cas13a enzyme is programmed with guide RNA to recognize genetic sequences of specific bacteria and resistance genes. When it finds its target, it produces a fluorescent signal that the fluorometer reads.

What the results showed
Of 306 assays, 274 (89.5%) fully or partly matched the gold-standard results for species and resistance genes, including 255 with perfect agreement. An earlier preprint described the platform as a low-cost tool for resource-limited settings. broadinstitutenih

Why it matters
Resistance to last-line antibiotics is a growing threat, and many hospitals can’t afford advanced molecular analyzers. A test that needs only simple equipment could help clinicians choose the right antibiotic sooner.

Limits to keep in mind

• The test works on blood cultures that have already turned positive, so culture is still the first step.
• It detects a fixed panel of species and genes, so it can’t replace full susceptibility testing.
• This is a research study, and it will need further validation before routine clinical use.

What it means for laboratories
Any new rapid method must be verified before use. Under ISO 15189 and ISO 17025, that means documented performance data, controls and acceptance criteria. Faster tests only help patients if the quality system around them is just as solid.

References

• Broad Institute publication record: BADLOCK, Science Advances, 2026; 12(40): eaeb6630. https://www.broadinstitute.org/publications/broad1382821
• Preprint: BADLOCK: A Rapid, Portable, Inexpensive Diagnostic for Bacterial Pathogen and Resistance Detection in Resource-Limited Settings. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12363746/