A test that searches for hundreds of chest germs at once can now report within a working day

DOCTORS TREATING a patient with a chest infection usually rely on two kinds of test. One grows germs from a sample on a dish, and the other, called PCR, looks for the genes of a fixed list of named viruses and bacteria. Both find only what they are set up to find, and between 15% and 30% of respiratory infections go undiagnosed, according to a new preprint from researchers at Great Ormond Street Hospital, University College London (UCL) and the University of Oxford. Their paper describes a way to run a much broader test within a working day.

The broadest approach is to read the genetic material of everything in a sample and then sort out which germs it came from. This is called metagenomics. The difficulty is that most of the genetic material in a sample from the nose or lungs belongs to the patient, so the germs make up a small share of what gets read. A refinement called hybrid capture deals with this by adding short strands of genetic material, known as probes, that match the genes of chosen germs. The probes bind to those germs' genes and pull them out of the sample before it is sequenced. The paper says this makes the test between 10 and 100 times more sensitive than reading everything. The team started from a commercial kit made by Illumina, a maker of sequencing machines, that targets more than 280 viruses, bacteria and fungi along with genes that let bacteria resist antibiotics.

The obstacle has been turnaround time, the time from taking a sample to reporting a result. The standard version of the kit needs about 11.4 hours of laboratory work, and adding up the step times the paper reports, including roughly 5.5 hours of sequencing and an hour of analysis, gives about 18 hours in all. The team shortened almost every step. A different extraction kit saved about 40 minutes, a faster method for copying viral RNA into DNA saved nearly two hours, and the time the probes are left to bind fell from 90 minutes to between 30 and 60 minutes. Their second fast version, which used an updated kit and a quicker sequencing run, needed about 6.6 hours of laboratory work and under ten hours from sample to result.

Speed is useful only if the test still finds the germs. The researchers checked it against 35 stored samples from patients at UCL hospitals and Great Ormond Street, compared with the germs that routine hospital tests had already found in them. In a first set of 12 samples, the standard and fast versions each found 11 of 15 such germs. In a further 23 samples, the second fast version found 19 of 24 and the first found 18 of 24. These differences were too small to show that any version was more or less sensitive than another. All versions also found germs that routine tests had missed, among them influenza B, rhinoviruses and Streptococcus pneumoniae, a common cause of pneumonia, which other tests later confirmed. Two hits for a type of adenovirus were probably contamination in the laboratory, and the fast versions read DNA germs such as cytomegalovirus more weakly than the standard one.

Two independent studies show why the time matters. In an intensive-care unit in Taiwan, doctors using the same commercial kit waited a median of 6.5 days for results, so the authors of a study in Critical Care described it mainly as an add-on to standard tests. At Guy's and St Thomas' hospitals in London, a pilot of a different sequencing method, published in 2024, tested 128 samples from 87 patients, reported results in a median 6.7 hours, and the authors say 48% of its results led doctors to change antibiotics. That pilot also shows that another method already works at a similar speed.

The new work has clear limits. It is a preprint that has not yet been reviewed by other scientists, and it rests on 35 hospital samples from one center. Illumina supplied the sequencing and capture reagents, and the authors declared no other competing interests. The study measured only whether the test detected germs, not whether faster results changed how patients were treated. The next step is to test whether results this fast change treatment for patients, in more samples and more hospitals.