Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: information science

‘Mesoscale’ swimmers could pave way for drug delivery robots inside the body

In physics, the mesoscale lies between the microscopic and the macroscopic. It is not just the domain of tiny living creatures like small larvae, shrimp, and jellyfish, but also where physics equations become extreme. While the macroscopic realm is governed by inertia and the microscopic by viscosity, the mesoscale is both and neither, requiring a new set of physics to describe it.

Now, physicists at Aalto University’s Department of Applied Physics have discovered how organisms swim in the mesoscale mix of viscosity and inertia. The study was recently published in the journal Communications Physics.

Led by Assistant Professor Matilda Backholm, the multidisciplinary team found the key to efficient swimming in this realm is not just moving faster or growing bigger, but a phenomenon of non-reciprocal motion known as time reversal symmetry breaking. The results help fill a knowledge gap in fundamental physics and could pave the way for applications such as mesorobotics; tiny robots injected inside a patient’s body for drug delivery or carrying out medical procedures.

Mapping 3D-super-enhancers with machine learning to pinpoint regulators of cell identity

Scientists usually study the molecular machinery that controls gene expression from the perspective of a linear, two-dimensional genome—even though DNA and its bound proteins function in three dimensions (3D). To better understand how key components of this machinery, such as super-enhancers, regulate genes in this 3D reality, scientists at St. Jude Children’s Research Hospital have developed a new algorithm called BOUQUET.

Using machine learning, BOUQUET reveals that sets of genes and their regulatory elements can interact within protein condensates, high-density membraneless droplets, in cells’ nuclei. The findings, which provide new insight into how cells regulate the genes that control their specialized identities, were published today in Nucleic Acids Research.

Cells express certain sets of genes to carry out specific functions; for example, a blood cell and a brain cell express different context-specific genes. There are 3 billion base pairs of human DNA, and the genes involved in cell identity are scattered throughout. Even more challenging, enhancers, DNA elements that activate gene expression, can be thousands of DNA bases away from their target genes.

Nonlinear photonic neuromorphic chips for spiking reinforcement learning

Photonic computing chips have made significant progress in accelerating linear computations, but nonlinear computations are usually implemented in the digital domain, which introduces additional system latency and power consumption, and hinders the implementation of fully functional photonic neural network chips. Here, we propose and fabricate a 16-channel programmable incoherent photonic neuromorphic computing chip by co-designing a simplified Mach–Zehnder interferometer (MZI) mesh and distributed feedback lasers with saturable absorber (DFBs-SA) array using different materials, enabling implementation of both linear and nonlinear spike computations in the optical domain through two separate chips. Furthermore, previous studies mainly focused on supervised learning and simple image classification tasks. Here, we propose a photonic spiking reinforcement learning (RL) architecture for the first, to our knowledge, time, and develop a software–hardware collaborative training-inference framework (in situ photonic training and hardware-aware fine-tuning) to address the challenge of training spiking RL models. We achieve large-scale, energy-efficient (photonic linear computation: 1.39 TOPS/W, photonic nonlinear computation: 987.65 GOPS/W), and low-latency (on-chip 320 ps) deployment of an entire layer of photonic spiking RL. Two RL benchmarks including the discrete CartPole task and the continuous Pendulum task are demonstrated experimentally based on the spiking proximal policy optimization (PPO) algorithm. The hardware–software collaborative computing reward value converges to 200 (−250) for the CartPole (Pendulum) tasks, respectively, comparable to that of a traditional PPO algorithm. This experimental demonstration addresses the challenge of the absence of large-scale on-chip photonic nonlinear spike computation and spiking RL training difficulty, and presents a high-speed and low-latency photonic spiking RL solution with promising application prospects in fields such as robot control, autonomous driving, and embodied intelligence.

The Immune Cell Atlas of “Longevity Molecular Tag”: Identification of Principal Immune Cell Subsets and Their Underlying Molecular Regulatory Mechanisms

Immunosenescence represents a critical aspect of the aging process. Centenarians, serving as a nature model of “healthy aging,” demonstrate a distinctive immune “compensatory adaptation” mechanism that contributes to the maintenance of immune homeostasis. However, the specific immune cell subsets involved and the molecular mechanisms underlying these phenotypic traits remain incompletely understood. In this study, we integrated single-cell RNA sequencing data spanning the entire lifespan of East Asian populations with bulk transcriptomic data from a centenarian cohort in Guangxi. Utilizing the Scissor algorithm, we identified immune cell subpopulations positively (Scissor+) and negatively (Scissor−) associated with longevity phenotypes, thereby constructing an immune cell atlas of “Longevity Molecular Tag.” Our findings indicate that Scissor+ cells predominantly comprise natural killer (NK) cells, CD8+ T cells, and γδ T cells, characterized by enhanced cytotoxic and immunomodulatory functions. Conversely, Scissor− cells mainly include CD4+ T cells, B cells, and dendritic cells (DCs), which are linked to inflammatory signaling pathways and Th17/Th1 differentiation. Trajectory analysis elucidated the differentiation pathways of NK, CD8+ T cells, CD4+ T cells, and B cells. Differentially expressed genes were enriched in pathways such as NF-κB signaling, T cell receptor signaling, and NK cell cytotoxicity. Furthermore, co-localization analysis revealed five eQTL-colocalized events (rs3793537–GLIPR2/CD72/TLN1 and rs8019902–TRDV2/TRDC) associated with longevity. Collectively, these results suggest that centenarians achieve immune equilibrium by remodeling cytotoxic immune lineages and finely tuning inflammatory responses, thereby promoting health span and longevity. This study offers novel insights into potential strategies for modulating immunosenescence.

The Multifaceted Paradigm of Rectal Cancer

“In a world where trimodality therapy has been the standard of care for so long, it’s remarkable to think that some of these cancers can be cured with a single systemic agent alone.”

🔗

The world of cancer treatment is a rapidly evolving creature, and rectal cancer is no exception. In particular, locally advanced rectal cancer has a number of valid treatment options. While it’s traditionally a surgical disease, in some cases we now have evidence for watch-and-wait approaches that spare patients the morbidity and toxicity associated with oncologic resections. But even when the goal is to get the patient to a total mesorectal excision (TME), several nuances can influence decision making. Suddenly, talking to a patient about rectal cancer has become as lengthy a discussion as those we have with intermediate-risk prostate cancer patients.

We currently have good evidence to suggest that total neoadjuvant therapy (TNT) should be standard of care for locally advanced rectal cancers. But even within this algorithm of chemotherapy and chemoradiation followed by surgery, questions abound. Which treatment should we start with? Which chemotherapy should be used? What radiation fractionation should we employ? And which concurrent chemotherapy should be paired with radiation? While the 5-year follow-up of the RAPIDO trial demonstrated a statistically significant increase in the locoregional recurrence rate (10% vs 6%) with short course radiation,1 this must be viewed through a critical lens, given that the two arms did not directly compare short-and long-course radiation. Perhaps it was the addition of neoadjuvant chemotherapy, delaying surgery, that resulted in a detriment to the locoregional control. Thus, short-course radiation is still indicated as a reasonable treatment option per NCCN guidelines.

What’s going on inside quantum computers? New method simplifies process tomography

Quantum computers work by applying quantum operations, such as quantum gates, to delicate quantum states. Ideally, quantum computers can solve complex equations at staggeringly fast speeds that vastly outpace regular computers. In real hardware, the operations of quantum computers often deviate from the ideal behavior because of device imperfections and unwanted noise from the environment. To build reliable quantum machines, researchers need a way to accurately determine what a quantum device is actually doing.

Quantum process tomography (QPT) is a standard method for this. However, traditional QPT becomes very costly as the system grows, because the number of required measurements and calculations increases rapidly with the number of qubits.

To address this challenge, a research team from Tohoku University, the Nara Institute of Science and Technology (NAIST), and the University of Information Technology (Vietnam National University, Ho Chi Minh City) has introduced a new framework called compilation-based quantum process tomography (CQPT). The work is published in Advanced Quantum Technologies.

All-in-Focus Fourier Ptychographic Microscopy via 3D Implicit Neural Representation

Microscopy has long been essential to biomedical research, enabling detailed analyses of complex samples. Fourier ptychographic microscopy (FPM), a computational imaging technique, provides high-resolution, wide-field images without requiring extensive hardware modifications. However, current FPM algorithms struggle with samples exhibiting depth variations, such as tilted or 3-dimensional (3D) objects. The limited depth of field (DoF) leads to images with only focal-plane areas in sharp focus, while regions outside appear blurred. To address this limitation, we propose an all-in-focus FPM algorithm using physics-informed 3D neural representations to reconstruct sharp, wide-field images of 3D objects under limited DoF. Unlike previous methods, our approach samples the full depth range to create a 3D feature volume that incorporates spatial and depth information.

/* */