ELD: transforming tissue analysis with next-gen landmark detection
A team of SciLifeLab researchers focusing on gene technology has recently published a study on a new groundbreaking method called Effortless Landmark Detection (ELD). By using ELD you can assemble a large amount of tissue sections into 3D structures, solving challenges in histology and spatially resolved transcriptomics.
In biological research, spatial landmarks are indispensable for navigating through tissue examinations, enabling scientists to compare samples, monitor specific areas, and accurately align tissue sections. Traditional methods for identifying these landmarks manually have become increasingly untenable with the advent of 3D reconstructions and high-throughput experiments that produce hundreds of samples.
“To address this challenge, we came up with a groundbreaking, unsupervised method for landmark detection and alignment named Effortless Landmark Detection (ELD). ELD leverages neural networks and a thin-plate splines alignment technique, specifically engineered to tackle the unique challenges presented by histology and spatially resolved transcriptomics,” says Markus Ekvall, PhD student at the Royal Institute of Technology (KTH).
This method not only ensures the stability of the detected landmarks but also improves the accuracy of tissue alignment. ELD simplifies the analysis of tissue samples, fostering a streamlined way of doing tissue analysis and speeding up scientific innovation.
“It’s about creating a spatial reference to compare various spatial experiments (gene expression, mass spectrometry imaging, morphology) that always differ slightly when working with individual tissue sections from a sample. The most impressive thing about the method is that it allows you to assemble hundreds of tissue sections into 3D structures,” says Joakim Lundeberg, professor at KTH.
