Trans-channel fluorescence learning improves high-content screening for Alzheimer’s disease therapeutics

Abstract

In microscopy-based drug screens, fluorescent markers carry critical information on how compounds affect different biological processes. However, practical considerations may hinder the use of certain fluorescent markers. Here, we present a deep learning method for overcoming this limitation. We accurately generated predicted fluorescent signals from other related markers and validated this new machine learning (ML) method on two biologically distinct datasets. We used the ML method to improve the selection of biologically active compounds for Alzheimer's disease (AD) from high-content high-throughput screening (HCS). The ML method identified novel compounds that effectively blocked tau aggregation, which would have been missed by traditional screening approaches unguided by ML. The method improved triaging efficiency of compound rankings over conventional rankings by raw image channels. We reproduced this ML pipeline on a biologically independent cancer-based dataset, demonstrating its generalizability.The approach is disease-agnostic and applicable across diverse fluorescence microscopy datasets..

Screening Techniques for Drug Discovery in Alzheimer’s Disease

“…This investigation was predominately conducted for identifying anti-SARS-CoV-2 molecules; however, one of the training protein target was BACE, which is a key target in AD. Deep learning using fluorescent markers have shown to improve the screening of small molecules in HCS . A three-channel data set, 4′,6-diamidino-2-phenylindole (DAPI) nuclear marker, YFP-tau, and AT8-pTau, was used to train the model and an archival HCS data set was applied.…”

“…Deep learning using fluorescent markers have shown to improve the screening of small molecules in HCS. 77 A three-channel data set, 4′,6-diamidino-2-phenylindole (DAPI) nuclear marker, YFPtau, and AT8-pTau, was used to train the model and an archival HCS data set was applied. The machine leaning method was able to identify compounds that can inhibit tau aggregation which are generally overlooked in traditional screening.…”

CellDeathPred: a deep learning framework for ferroptosis and apoptosis prediction based on cell painting

“…One example shows an image-to-image translation architecture for synthesizing three different fluorescence images from bright-field microscopy images to observe dead cells, nuclei, and cytoplasm of cells [14]. Another study proposed a U-Net architecture to synthesize AT8-pTau image given two DAPI and YFP-tau image channels [15]. With the potential of DL architectures in extracting meaningful features directly from microscopic images, recent studies proposed selfsupervised learning frameworks, including a framework for studying the temporal drug effect on cancer cell images, or a framework to learn phenotypic embeddings of HCS images using self-supervised triplet network [16,17].…”

Microsnoop: A Generalized Tool for Unbiased Representation of Diverse Microscopy Images

“…This approach involves applying a transformation step to the original images and training the model to learn the mapping between the transformed and the original image. Transformation can take various forms, such as a direct copy 19 , partial channel drop 20 , or image masking 21 , with masked visual representation learning being a particularly popular method in natural image studies 22-24 . Furthermore, recent advances in cell segmentation algorithms have indicated that networks trained on generalized data can possess remarkable generalization ability 25-27 .…”