Project example: image processing for digital PCR

We have been developing software for laboratories and research for many years. One example is a three-year research project that began in 2021. The aim was a new platform for digital multiplex assays. The instrument is intended to analyse a large number of DNA segments in a sample simultaneously, faster and more cheaply than before. It is based on digital PCR.

The polymerase chain reaction (PCR)

PCR is a standard method in molecular biology for amplifying DNA. Over many cycles, the enzyme DNA polymerase copies a target DNA segment. Each copy serves as a template again in the next cycle, hence the name chain reaction. Since its invention in 1983, PCR has been part of everyday work in research and laboratories, for example for detecting hereditary diseases or viruses.

Quantitative PCR: measuring against references

In quantitative real-time PCR, the instrument measures fluorescence during the reaction. Labelled probes light up as soon as they find the target DNA. The more DNA is produced, the brighter the sample becomes. The method became widely known through testing during the COVID-19 pandemic. The quantity is derived from comparison with reference samples. It is therefore only a relative measurement.

Digital PCR: counting instead of comparing

In digital PCR, the sample is divided into thousands of tiny droplets or chambers before the reaction. The reaction takes place separately in each one. Afterwards, every droplet is read out. If it lights up, it contained the target DNA. From the number of bright and dark droplets, the amount of DNA in the whole sample can be calculated statistically. Without reference samples and very precisely.

Digital PCR is less sensitive to substances that inhibit the reaction and also detects very small quantities. So far, however, the instruments have been considerably more expensive than those for quantitative PCR. This is where the project came in.

Where image processing helps

Reading out thousands of droplets or chambers individually is a typical task for image processing. The software finds every droplet in the image, measures its fluorescence, separates bright from dark ones and counts them. Multiplex assays add several colours, which it must distinguish cleanly. The more reliable this analysis, the more accurate the result.

More on this topic

Get in touch

Are you developing a laboratory instrument or analysing laboratory images? Briefly describe your task. We will get back to you quickly and advise you without obligation. Call +49 761 28522110 or use the contact form.

CONTACT

Select your language

First meetingarrange nowFirst meetingarrange now