Editorial

Advanced age is the major risk factor for idiopathic Parkinson's disease (PD), but to date the biological relationship between PD and ageing remains elusive. Here we describe the rationale and the design of the H2020 funded project "PROPAG-AGEING", whose aim is to characterize the contribution of the ageing process to PD development. We summarize current evidences that support the existence of a continuum between ageing and PD and justify the use of a Geroscience approach to study PD. We focus in particular on the role of inflammaging, the chronic, low-grade inflammation characteristic of elderly physiology, which can propagate and transmit both locally and systemically. We then describe PROPAG-AGEING design, which is based on the multi-omic characterization of peripheral samples from clinically characterized drug-naïve and advanced PD, PD discordant twins, healthy controls and "super-controls", i.e. centenarians, who never showed clinical signs of motor disability, and their offspring. Omic results are then validated in a large number of samples, including in vitro models of dopaminergic neurons and healthy siblings of PD patients, who are at higher risk of developing PD, with the final aim of identifying the molecular perturbations that can deviate the trajectories of healthy ageing towards PD development.

A prodromal phase of Parkinson's disease (PD) may precede motor manifestations by decades. PD patients' siblings are at higher risk for PD, but the prevalence and distribution of prodromal symptoms are unknown. The study objectives were (1) to assess motor and non-motor features estimating prodromal PD probability in PD siblings recruited within the European PROPAG-AGEING project; (2) to compare motor and non-motor symptoms to the well-established DeNoPa cohort. 340 PD siblings from three sites (Bologna, Seville, Kassel/Goettingen) underwent clinical and neurological evaluations of PD markers. The German part of the cohort was compared with German de novo PD patients (dnPDs) and healthy controls (CTRs) from DeNoPa. Fifteen (4.4%) siblings presented with subtle signs of motor impairment, with MDS-UPDRS-III scores not clinically different from CTRs. Symptoms of orthostatic hypotension were present in 47 siblings (13.8%), no different to CTRs (p = 0.072). No differences were found for olfaction and overall cognition; German-siblings performed worse than CTRs in visuospatial-executive and language tasks. 3/147 siblings had video-polysomnography-confirmed REM sleep behavior disorder (RBD), none was positive on the RBD Screening Questionnaire. 173/300 siblings had <1% probability of having prodromal PD; 100 between 1 and 10%, 26 siblings between 10 and 80%, one fulfilled the criteria for prodromal PD. According to the current analysis, we cannot confirm the increased risk of PD siblings for prodromal PD. Siblings showed a heterogeneous distribution of prodromal PD markers and probability. Additional parameters, including strong disease markers, should be investigated to verify if these results depend on validity and sensitivity of prodromal PD criteria, or if siblings' risk is not elevated.

Supervised link prediction aims at finding missing links in a network by learning directly from the data suitable criteria for classifying link types into existent or non-existent. Recently, along this line, subgraph-based methods learning a function that maps subgraph patterns to link existence have witnessed great successes. However, these approaches still have drawbacks. First, the construction of the subgraph relies on an arbitrary nodes selection, often ineffective. Second, the inability of such approaches to evaluate adaptively nodes importance reduces flexibility in nodes features aggregation, an important step in subgraph classification. To address these issues, a novel graph-classification based link-prediction model is proposed: Attention and Re-weighting based subgraph Classification for Link prediction (ARCLink). ARCLink first extracts a subgraph around the two nodes whose link should be predicted, by network reweighting, i.e. attributing a weight in the range 0-1 to all links of the original network, and then learns a function to map the subgraph to a continuous vector for classification, thus revealing the nature (non-existence/existence) of the unknown link. For leaning the mapping function, ARCLink generates a vector representation of the extracted subgraph by hierarchically aggregating nodes features according to nodes importance. In contrast to previous studies that either fully ignore or use fixed schemes to compute nodes importance, ARCLink instead learns nodes importance adaptively by employing attention mechanism. Through extensive experiments, ARCLink was validated on a series of real-world networks against state-of-the-art link prediction methods, consistently demonstrating its superior performances.

Methylage is an epigenetic marker of biological age that exploits the correlation between the methylation state of specific CG dinucleotides (CpGs) and chronological age (in years), gestational age (in weeks), cellular age (in cell cycles or as telomere length, in kilobases). Using DNA methylation data, methylage is measurable via the so called epigenetic clocks. Importantly, alterations of the correlation between methylage and age (age acceleration or deceleration) have been stably associated with pathological states and occur long before clinical signs of diseases become overt, making epigenetic clocks a potentially disruptive tool in preventive, diagnostic and also in forensic applications. Nevertheless, methylage dependency from CpGs selection, mathematical modelling, tissue specificity and age range, still makes the potential of this biomarker limited. In order to enhance model comparisons, interchange, availability, robustness and standardization, we organized a selected set of clocks within a hub webservice, EstimAge (Estimate of methylation Age, http://estimage.iac.rm.cnr.it), which intuitively and informatively enables quick identification, computation and comparison of available clocks, with the support of standard statistics.

Personalized medicine (PM) moves at the same pace of data and technology and calls for important changes in healthcare. New players are participating, providing impulse to PM. We review the conceptual foundations for PM and personalized healthcare and their evolution through scientific publications where a clear definition and the features of the different formulations are identifiable. We then examined PM policy documents of the International Consortium for Personalised Medicine and related initiatives to understand how PM stakeholders have been changing. Regional authorities and stakeholders have joined the race to deliver personalized care and are driving toward what could be termed as the next personalized healthcare. Their role as a key stakeholder in PM is expected to be pivotal.

Network embedding, also known as network representation learning, aims at defining low-dimensional, continuous vector representation of nodes to maximally preserve the network structure. Recent efforts attempt to extend network embedding to attributed networks where nodes are enriched with descriptors, to enhance interpretability. However, most of these efforts seldom consider the additional knowledge relevant to the aim of the downstream network analysis, i.e. task-related information. When they do, they are analysis-specific and thus lack adaptability to alternative tasks. In this article, a unified framework TANE is proposed to learn Task-oriented Attributed Network Embedding that jointly, maximally and consistently preserves multiple types of network information to generate rich nodes representations, robust to a variety of analyses. The framework can flexibly adapt to, and be readily modified for, different network-based tasks in an end-to-end way. The results of extensive experiments on well-known and commonly used datasets demonstrate that the proposed framework TANE can achieve superior performance over state-of-the-art methods in two commonly performed tasks: node classification and link prediction.

Motivation Inflammation is part of the complex function that addresses harmful stimuli, and the first phase of wound healing (WH), which guarantees living systems' homeostasis. Deviances from physiology make inflammation turn acute (sepsis, 11M death/y) or chronic (non-communicable diseases, 41M death/y). Therefore, tackling inflammation is a key priority. We recently proposed (Maturo et al., 2020) to revise the conventional inflammatory pathway (innate immune response) to include WH (expanded inflammatory pathway). Methods We manually identified the Reactome pathways that include all reactions and species relevant to WH. Cytoscape was then used to perform the union of the SBML converted pathways, with the largest connected component being retained (732 nodes, 13.944 edges). The same was done for the innate immune response (R-HSA-168249.8) with 487 nodes, 11.744 edges. We then focused on: NF-kB (fundamental hub in all inflammatory reactions), TNF-? (renown target of inflammatory diseases) and RAC1 (key player in mechanotransduction events of WH). Results Preliminary topological results highlight the stability of closeness centrality, i.e. all molecules preserve their efficiency in spreading information. Conversely, betweenness centrality is stable for NF-kB (0.068), confirming NF-kB relevance, while halving its (very low) value in the expanded pathway for TNF-? (from 2.85E-06 to 1.29E-06). This indicates that the ability to bridge different parts of the graphs is less effective if we consider inflammation as an expanded concept, possibly contributing to explain the many side effects of anti-TNF-? therapies. Interestingly, RAC1 presents stable betweenness (from 0.094 to 0.093), comparable to NF-kB, supporting the hypothesis that WH-leveraging therapies could act on a relevant and stable target, so far neglected (Nardini et al., 2016).

The aim of the work is to propose a methodology for the stimulation of a 3D in vitro skin model to activate wound healing. The presented work is in the frame of the national research project, CronXCov, "Checking the CHRONIC to prevent COVid-19", devoted to understand how physiologic and inflamed skin on chip 3D models evolve upon a range of physical (e.g., electrical, mechanical, optical) stimulations, over time. Thanks to the 3D modelling, using Next Generation Sequencing and the network medicine frame of analysis to process the data, we will systematically characterize the effects of the applied stimuli, offering new insight for the exploitation of wound healing.

Background: High-throughput technologies enable the cost-effective collection and analysis of DNA methylation data throughout the human genome. This naturally entails missing values management that can complicate the analysis of the data. Several general and specific imputation methods are suitable for DNA methylation data. However, there are no detailed studies of their performances under different missing data mechanisms -(completely) at random or not- and different representations of DNA methylation levels (beta andM-value).

Meta-analysis is a powerful means for leveraging the hundreds of experiments being run worldwide into more statistically powerful analyses. This is also true for the analysis of omic data, including genome-wide DNA methylation. In particular, thousands of DNA methylation profiles generated using the Illumina 450k are stored in the publicly accessible Gene Expression Omnibus (GEO) repository. Often, however, the intensity values produced by the BeadChip (raw data) are not deposited, therefore only pre-processed values-obtained after computational manipulation-are available. Pre-processing is possibly different among studies and may then affect meta-analysis by introducing non-biological sources of variability. Introduction

The feasibility of liver transplantation from old healthy donors suggests that this organ is able to preserve its functionality during aging. To explore the biological basis of this phenomenon, we characterized the epigenetic profile of liver biopsies collected from 45 healthy liver donors ranging from 13 to 90 years old using the Infinium HumanMethylation450 BeadChip. The analysis indicates that a large remodeling in DNA methylation patterns occurs, with 8823 age-associated differentially methylated CpG probes. Notably, these age-associated changes tended to level off after the age of 60, as confirmed by Horvath's clock. Using stringent selection criteria we further identified a DNA methylation signature of aging liver including 75 genomic regions. We demonstrated that this signature is specific for liver compared to other tissues and that it is able to detect biological age-acceleration effects associated with obesity. Finally we combined DNA methylation measurements with available expression data. Although the intersection between the two omic characterizations was low, both approaches suggested a previously unappreciated role of epithelial-mesenchymal transition and Wnt signaling pathways in the aging of human liver.

Imputation of methylation data

Background and limitations Impaired wound healing (WH) and chronic inflammation are hallmarks of non-communicable diseases (NCDs). However, despite WH being a recognized player in NCDs, mainstream therapies focus on (un)targeted damping of the inflammatory response, leaving WH largely unaddressed, owing to three main factors. The first is the complexity of the pathway that links inflammation and wound healing; the second is the dual nature, local and systemic, of WH; and the third is the limited acknowledgement of genetic and contingent causes that disrupt physiologic progression of WH. Proposed approach Here, in the frame of Predictive, Preventive, and Personalized Medicine (PPPM), we integrate and revisit current literature to offer a novel systemic view on the cues that can impact on the fate (acute or chronic inflammation) of WH, beyond the compartmentalization of medical disciplines and with the support of advanced computational biology. Conclusions This shall open to a broader understanding of the causes for WH going awry, offering new operational criteria for patients' stratification (prediction and personalization). While this may also offer improved options for targeted prevention, we will envisage new therapeutic strategies to reboot and/or boost WH, to enable its progression across its physiological phases, the first of which is a transient acute inflammatory response versus the chronic low-grade inflammation characteristic of NCDs.

Results: We present a simple and computationally efficient imputation method, metyhLImp, based on linear regression. The rationale of the approach lies in the observation that methylation levels show a high degree of inter-sample correlation. We performed a comparative study of our approach with other imputation methods on DNA methylation data of healthy and disease samples from different tissues. Performances have been assessed both in terms of imputation accuracy and in terms of the impact imputed values have on mAge estimation. In comparison to existing methods, our linear regression model proves to perform equally or better and with good computational efficiency. The results of our analysis provide recommendations for accurate estimation of missing methylation values. Motivation: DNA methylation is a stable epigenetic mark with major implications in both physiological (development, aging) and pathological conditions (cancers and numerous diseases). Recent research involving methylation focuses on the development of molecular age estimation methods based on DNA methylation levels (mAge). An increasing number of studies indicate that divergences between mAge and chronological age may be associated to age-related diseases. Current advances in high-throughput technologies have allowed the characterization of DNA methylation levels throughout the human genome. However, experimental methylation profiles often contain multiple missing values that can affect the analysis of the data and also mAge estimation. Although several imputation methods exist, a major deficiency lies in the inability to cope with large datasets, such as DNA methylation chips. Specific methods for imputing missing methylation data are therefore needed.

Chronic inflammatory autoimmune disorders are systemic diseases with increasing incidence and still lack a cure. More recently, attention has been placed in understanding gastrointestinal (GI) dysbiosis and, although important progress has been made in this area, it is currently unclear to what extent microbiome manipulation can be used in the treatment of autoimmune disorders. Via the use of appropriate models, rheumatoid arthritis (RA), a well-known exemplar of such pathologies, can be exploited to shed light on the currently overlooked effects of existing therapies on the GI microbiome. In this direction, we here explore the crosstalk between the GI microbiome and the host immunity in model arthritis (collagen induced arthritis, CIA). By exploiting omics from samples of limited invasiveness (blood and stools), we assess the host-microbiome responses to standard therapy (methotrexate, MTX) combined with mechanical subcutaneous stimulation (MS) and to mechanical stimulation alone. When MS is involved, results reveal the sphingolipid metabolism as the trait d'union among known hallmarks of (model) RA, namely: Imbalance in the S1P-S1PR1 axis, expansion of Prevotella sp., and invariant Natural Killer T (iNKT)-penia, thus offering the base of a rationale to mechanically modulate this pathway as a therapeutic target in RA.

A growing amount of evidences indicates that inflammaging - the chronic, low grade inflammation state characteristic of the elderly - is the result of genetic as well as environmental or stochastic factors. Some of these, such as the accumulation of senescent cells that are persistent during aging or accompany its progression, seem to be sufficient to initiate the aging process and to fuel it. Others, like exposure to environmental compounds or infections, are temporary and resolve within a (relatively) short time. In both cases, however, a cellular memory of the event can be established by means of epigenetic modulation of the genome. In this review we will specifically discuss the relationship between epigenetics and inflammaging. In particular, we will show how age-associated epigenetic modifications concerned with heterochromatin loss and gene-specific remodelling, can promote inflammaging. Furthermore, we will recall how the exposure to specific nutritional, environmental and microbial stimuli can affect the rate of inflammaging through epigenetic mechanisms, touching also on the recent insight given by the concept of trained immunity.

In this review we analyse the impact of age-associated factors such as inflammaging, immunosenescence and immunobiography on immune response to vaccination in the elderly, and we consider systems biology approaches as a mean for integrating a multitude of data in order to rationally design vaccination approaches specifically tailored for the elderly. The unprecedented increase of life expectancy challenges society to protect the elderly from morbidity and mortality making vaccination a crucial mean to safeguard this population. Indeed, infectious diseases, such as influenza and pneumonia, are among the top killers of elderly people in the world. Elderly individuals are more prone to severe infections and less responsive to vaccination prevention, due to immunosenescence combined with the progressive increase of a proinflammatory status characteristic of the aging process (inflammaging). These factors are responsible for most age-related diseases and correlate with poor response to vaccination. Therefore, it is of utmost interest to deepen the knowledge regarding the role of inflammaging in vaccination responsiveness to support the development of effective vaccination strategies designed for elderly.

Under the current deluge of omics, module networks distinctively emerge as methods capable of not only identifying inherently coherent groups (modules), thus reducing dimensionality, but also hypothesizing cause-effect relationships between modules and their regulators. Module networks were first designed in the transcriptomic era and further exploited in the multi-omic context to assess (for example) miRNA regulation of gene expression. Despite a number of available implementations, expansion of module networks to other omics is constrained by a limited characterization of the solutions' (modules plus regulators) accuracy and stability-an immediate need for the better characterization of molecular biology complexity in silico. We hence carefully assessed for LemonTree-a popular and open source module network implementation-the dependency of the software performances (sensitivity, specificity, false discovery rate, solutions' stability) on the input parameters and on the data quality (sample size, expression noise) based on synthetic and real data. In the process, we uncovered and fixed an issue in the code for the regulator assignment procedure. We concluded this evaluation with a table of recommended parameter settings. Finally, we applied these recommended settings to gut-intestinal metagenomic data from rheumatoid arthritis patients, to characterize the evolution of the gut-intestinal microbiome under different pharmaceutical regimens (methotrexate and prednisone) and we inferred innovative clinical recommendations with therapeutic potential, based on the computed module network.

Motivation: Cells derived by cellular engineering, i.e. differentiation of induced pluripotent stem cells and direct lineage reprogramming, carry a tremendous potential for medical applications and in particular for regenerative therapies. These approaches consist in the definition of lineage-specific experimental protocols that, by manipulation of a limited number of biological cues-niche mimicking factors, (in) activation of transcription factors, to name a few-enforce the final expression of cell-specific (marker) molecules. To date, given the intricate complexity of biological pathways, these approaches still present imperfect reprogramming fidelity, with uncertain consequences on the functional properties of the resulting cells. Results: We propose a novel tool eegc to evaluate cellular engineering processes, in a systemic rather than marker-based fashion, by integrating transcriptome profiling and functional analysis. Our method clusters genes into categories representing different states of (trans) differentiation and further performs functional and gene regulatory network analyses for each of the categories of the engineered cells, thus offering practical indications on the potential lack of the reprogramming protocol.