Leveraging a large language model to predict protein phase transition: A physical, multiscale, and interpretable approach.

Protein phase transitions (PPTs) from the soluble state to a dense liquid phase (forming droplets via liquid-liquid phase separation) or to solid aggregates (such as amyloids) play key roles in pathological processes associated with age-related diseases such as Alzheimer's disease. Several computational frameworks are capable of separately predicting the formation of droplets or amyloid aggregates based on protein sequences, yet none have tackled the prediction of both within a unified framework. Recently, large language models (LLMs) have exhibited great success in protein structure prediction; however, they have not yet been used for PPTs. Here, we fine-tune a LLM for predicting PPTs and demonstrate its usage in evaluating how sequence variants affect PPTs, an operation useful for protein design. In addition, we show its superior performance compared to suitable classical benchmarks. Due to the "black-box" nature of the LLM, we also employ a classical random forest model along with biophysical features to facilitate interpretation. Finally, focusing on Alzheimer's disease-related proteins, we demonstrate that greater aggregation is associated with reduced gene expression in Alzheimer's disease, suggesting a natural defense mechanism.

More than bad luck: Cancer and aging are linked to replication-driven changes to the epigenome.

Aging is a leading risk factor for cancer. While it is proposed that age-related accumulation of somatic mutations drives this relationship, it is likely not the full story. We show that aging and cancer share a common epigenetic replication signature, which we modeled using DNA methylation from extensively passaged immortalized human cells in vitro and tested on clinical tissues. This signature, termed CellDRIFT, increased with age across multiple tissues, distinguished tumor from normal tissue, was escalated in normal breast tissue from cancer patients, and was transiently reset upon reprogramming. In addition, within-person tissue differences were correlated with predicted lifetime tissue-specific stem cell divisions and tissue-specific cancer risk. Our findings suggest that age-related replication may drive epigenetic changes in cells and could push them toward a more tumorigenic state.

Detecting Sensitive Mobility Features for Parkinson's Disease Stages Via Machine Learning.

BACKGROUND: It is not clear how specific gait measures reflect disease severity across the disease spectrum in Parkinson's disease (PD). OBJECTIVE: To identify the gait and mobility measures that are most sensitive and reflective of PD motor stages and determine the optimal sensor location in each disease stage. METHODS: Cross-sectional wearable-sensor records were collected in 332 patients with PD (Hoehn and Yahr scale I-III) and 100 age-matched healthy controls. Sensors were adhered to the participant's lower back, bilateral ankles, and wrists. Study participants walked in a ~15-meter corridor for 1 minute under two walking conditions: (1) preferred, usual walking speed and (2) walking while engaging in a cognitive task (dual-task). A subgroup (n = 303, 67% PD) also performed the Timed Up and Go test. Multiple machine-learning feature selection and classification algorithms were applied to discriminate between controls and PD and between the different PD severity stages. RESULTS: High discriminatory values were found between motor disease stages with mean sensitivity in the range 72%-83%, specificity 69%-80%, and area under the curve (AUC) 0.76-0.90. Measures from upper-limb sensors best discriminated controls from early PD, turning measures obtained from the trunk sensor were prominent in mid-stage PD, and stride timing and regularity were discriminative in more advanced stages. CONCLUSIONS: Applying machine-learning to multiple, wearable-derived features reveals that different measures of gait and mobility are associated with and discriminate distinct stages of PD. These disparate feature sets can augment the objective monitoring of disease progression and may be useful for cohort selection and power analyses in clinical trials of PD. © 2021 International Parkinson and Movement Disorder Society.

The biomechanical efficacy of dressings in preventing heel ulcers.

The heels are the most common site for facility-acquired pressure ulcers (PUs), and are also the most susceptible location for deep tissue injuries. The use of multilayer prophylactic dressings to prevent heel PUs is a relatively new prevention concept, generally aimed at minimizing the risk for heel ulcers (HUs) through mechanical cushioning and reduction of friction at the dressing-support interface. We used 9 finite element model variants of the posterior heel in order to evaluate the biomechanical performance of a multilayer dressing in prevention of HUs during supine lying. We compared volumetric exposures of the loaded soft tissues to effective and maximal shear strains, as well as peak stresses in the Achilles tendon, without any dressing and with a single-layer or a multilayer dressing (Mepilex(®) Border Heel-type), on supports with different stiffnesses. The use of the multilayer dressing consistently and considerably reduced soft tissue exposures to elevated strains at the posterior heel, on all of the tested support surfaces and when loaded with either pure compression or combined compression and shear. The aforementioned multilayer design showed (i) clear benefit over a single-layer dressing in terms of dissipating tissue strains, by promoting internal shear in the dressing which diverts loads from tissues; (ii) a protective effect that was consistent on supports with different stiffnesses. Recent randomized controlled trials confirmed the efficacy of the simulated multilayer dressing, and so, taken together with this modeling work, the use of a prophylactic multilayer dressing indicates a great promise in taking this route for prevention.

Effects of accumulation of lipid droplets on load transfer between and within adipocytes.

Adipogenesis, a process of cell proliferation followed by the accumulation of lipid droplets (LDs), is accompanied by morphological changes in adipocytes, leading to a gradual rise in the structural stiffness of these cells. The increase in cellular structural stiffness can potentially influence the localized deformations of adjacent adipocytes in weight-bearing fat tissues, which, based on previous work, may accelerate intracytoplasmatic lipid production to form even larger and more tightly packed intracellular LDs. This process is based on mechanotransduction phenomena which are hypothesized (again, following empirical studies), to play a critical role in "en mass" adipocyte hypertrophy, and hence are important to characterize through computational modeling. Accordingly, we examined here how maturing adipocytes may affect localized loads acting on adjacent immature cells, using a set of finite element models of adipocytes embedded in an extracellular matrix. The peak strain energy density at the plasma membrane (PM) of the adipocytes, when constructs were externally loaded, was found to depend on the levels of lipid accumulation in the neighboring cells if the external compressive and shear deformations were large enough ([Formula: see text] and [Formula: see text], respectively). The mechanosignaling transduces through the PM and could therefore affect intracellular pathways to produce more lipid contents. Our results support the theory of deformation-induced differentiation in adipocytes. The findings are thus relevant in the context of a sedentary lifestyle, in which sustained deformations of weight-bearing adipose tissues may activate a positive feedback loop that promotes the "en mass" differentiation of cells, which subsequently increases the total mass of living fat tissues.