Search

Background: Overt sentence reading in mild cognitive impairment (MCI) and mild-to-moderate Alzheimer’s disease (AD) has been associated with slowness of speech, characterized by a higher number of pauses, shorter speech units and slower speech rate and attributed to reduced working memory/ attention and language capacity. Objective: This preliminary case-control study investigates whether the temporal organization of speech is associated with the volume of brain regions involved in overt sentence reading and explores the discriminative ability of temporal speech parameters and standard volumetric MRI measures for the classification of MCI and AD. Method: Individuals with MCI, mild-to-moderate AD, and healthy controls (HC) had a structural MRI scan and read aloud sentences varying in cognitive-linguistic demand (length). The association between speech features and regional brain volumes was examined by linear mixed-effect modeling. Genetic programming was used to explore the discriminative ability of temporal and MRI features. Results: Longer sentences, slower speech rate, higher number of pauses and shorter interpausal units were associated with reduced volumes of the reading network. Speech-based classifiers performed similarly to the MRI-based classifiers for MCI-HC (67% vs 68%) and slightly better for AD-HC (80% vs 64%) and AD-MCI (82% vs 59%). Adding the speech features to the MRI features slightly improved performance of MRI-based classification for AD-HC and MCI-HC but not HC-MCI. Conclusion: The temporal organization of speech in overt sentence reading reflects underlying volume reductions. It may represent a sensitive marker for early assessment of structural changes and cognitive-linguistic deficits associated with healthy aging, MCI, and AD.

BackgroundEvidence is limited regarding the cumulative effect of risk factors on cognitive decline and the added value of physical function for cognitive function trajectory stratification. We operationalize 13 modifiable dementia risk factors in a scoring system and investigate the relationship between this brain health score, combined with simple measures of physical function, and risk of cognitive decline.MethodsPopulation-based cohort study of persons aged 50 and older from the Irish Longitudinal Study on Ageing without a history of dementia at baseline who underwent repeated neuropsychological tests (8.08 ± 0.3-year follow-up) were included in the analyses. Exposures were the number of brain health metrics (defined by the Lancet Commission on Dementia Prevention, Intervention, and Care report) at recommended optimal levels. Physical function exposures included Timed Up and Go, dual-task walking speed, and grip strength. Each health metric and physical function measure at the recommended level was assigned a value of 1 and combined to generate brain health scores. Relationship with group-based trajectories of global cognitive function (multidomains composite score), estimated using K-means for longitudinal data, was assessed via ordinal logistic regressions.ResultsAmong 2 327 participants (mean age, 61 years; 54% women), each additional optimal metric on the brain health score (odds 0.67 [0.62, 0.73]) was associated with reduced odds of cognitive decline. Adding Timed Up and Go (odds 0.71 [0.59, 0.84]) and dual-task walking speed (odds 0.74 [0.63, 0.89]) further improved model fit (ΔAIC = 14.8).ConclusionThese findings support the promotion and maintenance of physical function in addition to brain health strategies to reduce the risk of cognitive decline.

OBJECTIVE To identify serum protein biomarkers that might distinguish patients with early inflammatory arthritis (IA) with psoriatic arthritis (PsA) from those with rheumatoid arthritis (RA) and may be used to support appropriate early intervention. METHODS The serum proteome of patients with PsA and patients with RA was interrogated using nano-liquid chromatography mass spectrometry (nano-LC-MS/MS) (n = 64 patients), an aptamer-based assay (SomaScan) targeting 1,129 proteins (n = 36 patients), and a multiplexed antibody assay (Luminex) for 48 proteins (n = 64 patients). Multiple reaction monitoring (MRM) assays were developed to evaluate the performance of putative markers using the discovery cohort (n = 60 patients) and subsequently an independent cohort of PsA and RA patients (n = 167). RESULTS Multivariate machine learning analysis of the protein discovery data from the 3 platforms revealed that it was possible to differentiate PsA patients from RA patients with an area under the curve (AUC) of 0.94 for nano-LC-MS/MS, 0.69 for bead-based immunoassay measurements, and 0.73 for aptamer-based analysis. Subsequently, in the separate verification and evaluation studies, random forest models revealed that a subset of proteins measured by MRM could differentiate PsA and RA patients with AUCs of 0.79 and 0.85, respectively. CONCLUSION We present a serum protein biomarker panel that can separate patients with early-onset IA with PsA from those with RA. With continued evaluation and refinement using additional and larger patient cohorts, including those with other arthropathies, we suggest that the panel identified here could contribute to improved clinical decision making.

AIMS: There is a critical need for better biomarkers so that heart failure can be diagnosed at an earlier stage and with greater accuracy. The purpose of this study was to design a robust mass spectrometry (MS)-based assay for the simultaneous measurement of a panel of 35 candidate protein biomarkers of heart failure, in blood. The overall aim was to evaluate the potential clinical utility of this biomarker panel for prediction of heart failure in a cohort of 500 patients.METHODS AND RESULTS: Multiple reaction monitoring (MRM) MS assays were designed with Skyline and Spectrum Mill PeptideSelector software and developed using nanoflow reverse phase C18 chromatographic Chip Cube-based separation, coupled to a 6460 triple quadrupole mass spectrometer. Optimized MRM assays were applied, in a sample-blinded manner, to serum samples from a cohort of 500 patients with heart failure and non-heart failure (non-HF) controls who had cardiovascular risk factors. Both heart failure with reduced ejection fraction (HFrEF) patients and heart failure with preserved ejection fraction (HFpEF) patients were included in the study. Peptides for the Apolipoprotein AI (APOA1) protein were the most significantly differentially expressed between non-HF and heart failure patients (P = 0.013 and P = 0.046). Four proteins were significantly differentially expressed between non-HF and the specific subtypes of HF (HFrEF and HFpEF); Leucine-rich-alpha-2-glycoprotein (LRG1, P < 0.001), zinc-alpha-2-glycoprotein (P = 0.005), serum paraoxanse/arylesterase (P = 0.013), and APOA1 (P = 0.038). A statistical model found that combined measurements of the candidate biomarkers in addition to BNP were capable of correctly predicting heart failure with 83.17% accuracy and an area under the curve (AUC) of 0.90. This was a notable improvement on predictive capacity of BNP measurements alone, which achieved 77.1% accuracy and an AUC of 0.86 (P = 0.005). The protein peptides for LRG1, which contributed most significantly to model performance, were significantly associated with future new onset HF in the non-HF cohort [Peptide 1: odds ratio (OR) 2.345 95% confidence interval (CI) (1.456-3.775) P = 0.000; peptide 2: OR 2.264 95% CI (1.422-3.605), P = 0.001].CONCLUSIONS: This study has highlighted a number of promising candidate biomarkers for (i) diagnosis of heart failure and subtypes of heart failure and (ii) prediction of future new onset heart failure in patients with cardiovascular risk factors. Furthermore, this study demonstrates that multiplexed measurement of a combined biomarker signature that includes BNP is a more accurate predictor of heart failure than BNP alone.

The aging process is characterized by the presence of high interindividual variation between individuals of the same chronical age prompting a search for biomarkers that capture this heterogeneity. Epigenetic clocks measure changes in DNA methylation levels at specific CpG sites that are highly correlated with calendar age. The discrepancy resulting from the regression of DNA methylation age on calendar age is hypothesized to represent a measure of biological aging with a positive/negative residual signifying age acceleration (AA)/deceleration, respectively. The present study examines the associations of 4 epigenetic clocks—Horvath, Hannum, PhenoAge, GrimAge—with a wide range of clinical phenotypes (walking speed, grip strength, Fried frailty, polypharmacy, Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MOCA), Sustained Attention Reaction Time, 2-choice reaction time), and with all-cause mortality at up to 10-year follow-up, in a sample of 490 participants in the Irish Longitudinal Study on Ageing (TILDA). HorvathAA and HannumAA were not predictive of health; PhenoAgeAA was associated with 4/9 outcomes (walking speed, frailty MOCA, MMSE) in minimally adjusted models, but not when adjusted for other social and lifestyle factors. GrimAgeAA by contrast was associated with 8/9 outcomes (all except grip strength) in minimally adjusted models, and remained a significant predictor of walking speed, .polypharmacy, frailty, and mortality in fully adjusted models. Results indicate that the GrimAge clock represents a step-improvement in the predictive utility of the epigenetic clocks for identifying age-related decline in an array of clinical phenotypes promising to advance precision medicine.

Gait speed is a measure of general fitness. Changing from usual (UGS) to maximum (MGS) gait speed requires coordinated action of many body systems. Gait speed reserve (GSR) is defined as MGS–UGS. From a shortlist of 88 features across five categories including sociodemographic, cognitive, and physiological, we aimed to find and compare the sets of predictors that best describe UGS, MGS, and GSR. For this, we leveraged data from 3,925 adults aged 50+ from Wave 3 of The Irish Longitudinal Study on Ageing (TILDA). Features were selected by a histogram gradient boosting regression-based stepwise feature selection pipeline. Each model’s feature importance and input–output relationships were explored using TreeExplainer from the Shapely Additive Explanations explainable machine learning package. The mean Radj2 (SD) from fivefold cross-validation on training data and the Radj2 score on test data were 0.38 (0.04) and 0.41 for UGS, 0.45 (0.04) and 0.46 for MGS, and 0.19 (0.02) and 0.21 for GSR. Each model selected features across all categories. Features common to all models were age, grip strength, chair stands time, mean motor reaction time, and height. Exclusive to UGS and MGS were educational attainment, fear of falling, Montreal cognitive assessment errors, and orthostatic intolerance. Exclusive to MGS and GSR were body mass index (BMI), and number of medications. No features were selected exclusively for UGS and GSR. Features unique to UGS were resting-state pulse interval, Center for Epidemiologic Studies Depression Scale (CESD) depression, sit-to-stand difference in diastolic blood pressure, and left visual acuity. Unique to MGS were standard deviation in sustained attention to response task times, resting-state heart rate, smoking status, total heartbeat power during paced breathing, and visual acuity. Unique to GSR were accuracy proportion in a sound-induced flash illusion test, Mini-mental State Examination errors, and number of cardiovascular conditions. No interactions were present in the GSR model. The four features that overall gave the most impactful interactions in the UGS and MGS models were age, chair stands time, grip strength, and BMI. These findings may help provide new insights into the multisystem predictors of gait speed and gait speed reserve in older adults and support a network physiology approach to their study.

Multimorbidity (the presence of multiple medical conditions) is well known to increase with age. People with multimorbidities often have higher physical and functional decline as well as increased mortality. Despite growing evidence that integrated and collaborative care improves many undesirable outcomes of multimorbidity, the majority of health systems are based around treating individual diseases. A pattern analysis of comorbidities using network graphs and a novel use of association rules was conducted to investigate disease associations on 6101 Irish adults aged 50+. The complex network of morbidities and differences in the prevalence and interactions of these morbidities by sex was also assessed. Gender specific differences in disease prevalence was found for 22/31 medical conditions included in this study. Females had a more complex network of disease associations than males with strong associations found between arthritis, osteoporosis and thyroid issues among others. To assess the strength of these associations we provide probabilities of being diagnosed with a comorbid condition given the presence of an index morbidity for 639 pairwise combinations. This information can be used to guide clinicians in deciding which comorbidities should be incorporated into comprehensive assessments in addition to anticipating likely future morbidities and thus developing prevention strategies.

Background Slow gait speed and Timed Up-and Go (TUG) are often independent predictors of falls in regression analysis, although their ability to discriminate between fallers and non-fallers is questionable. Random forest plots can model complex interactions between predictor variables and have high classification accuracy. We compare the predictive and discriminative ability of UGS and TUG in predicting recurrent falls over 4 years using poisson regression and random forest plots. Methods Data from the first three waves of The Irish Longitudinal Study on Ageing (TILDA), a population-based study of community-dwelling adults aged ≥50 years were used. Baseline physical, neuro-cognitive, sensory and behavioural health were assessed. The outcome was recurrent (≥2) falls at Wave 2 or Wave 3. Poisson regression analysis was used to examine associations between UGS, TUG and falls adjusting for covariates. Random forest models were used. Predictive accuracy was calculated using 5 fold cross-validation and as there was class imbalance, the algorithm was trained using under-sampling of the larger class. Classification rate, area under the receiver operating characteristic curve (AUC) and area under the precision recall curve (PRROC) were obtained to assess predictive accuracy. Results In poisson regression analysis (n=4918), TUG predicted recurrent falls independent of covariates including UGS (IRR=1.06, 95% CI: 1.01, 1.12, p Conclusion Impaired mobility i.e. slower TUG performance, is an independent predictor of future recurrent falls, however it does not discriminate well between fallers and non-fallers. This is highlighted by the PRROC which provides a more conservative estimate of predictive accuracy than AUC as it accounts for the ability to identify both fallers and non-fallers. The analysis highlights the multifactorial nature and complexity of falls, and supports the need for comprehensive falls assessment.

Background In ageing, multisensory integration, i.e. the ability to combine efficiently information from different sensory modalities, is emerging as a stand-alone contributor to explaining cognitive and functional deficits. Experimental evidence shows that inefficient multisensory integration is associated with cognitive impairment, falls and balance maintenance; however these findings need to be supported by large population representative studies. We utilised the Sound-induced flash illusion (SIFI) as test of multisensory efficiency; the test was conducted on a population representative sample from the Irish Longitudinal Study on Ageing (TILDA). We hypothesized that Susceptibility to the SIFI would increase with ageing and would be associated with poorer Montreal Cognitive Assessment (MoCA) scores. Methods Participants: 3,955 adults aged 50 years and over. Participants provided informed consent. Data were drawn from the third wave of the TILDA study, in which participants took part in a Computer Assisted Interview and a Health Assessment. The SIFI test was part of the Health Assessment; SIFI is a computer-based test in which participants see a series of white dots appearing quickly on the screen, which participants are required to count; the dots (flashes) can be paired with 0, 1 or 2 beeps; when one dot is paired with two beeps, there is the illusory perception that two dots are presented. A hierarchical Bayesian, ordinal-regression model was used to determine which variables predicted audio-visual integration (SIFI proportion of correct responses, i.e. no illusions). We controlled for a range of covariates. Results As predicted, higher susceptibility to the SIFI, indicating higher integration of audio-visual information, was associated with older age, and poorer scores at MoCA. Female sex was also associated with higher susceptibility, which represents a novel result in the literature. Conclusion The present study presents the first findings on multisensory integration in a large population representative study. They confirm that inefficient integration is associated with ageing, poorer cognition and, unexpectedly, being female.

Gait speed is a measure of general fitness. Changing from usual (UGS) to maximum (MGS) gait speed requires a general effort across many body systems. The difference, MGS – UGS, is defined as gait speed reserve (GSR). In the present study, using 3925 participants aged 50+ from Wave 3 of The Irish Longitudinal Study on Ageing (TILDA), we used a gradient boosted trees-based stepwise feature selection pipeline for the discovery of clinically relevant predictors of GSR, UGS, and MGS using a shortlist of 88 features across 5 categories (socio-demographics/anthropometrics/medical history; cardiovascular system; physical strength; sensory; and cognitive/psychological). The TreeSHAP explainable machine learning package was used to analyse the input-output relationships of the three models.The meanfrom 5-fold cross validation on training data and thescore on test data for the models are: 0.38 (0.04) and 0.41 for UGS; 0.45 (0.04) and 0.46 for MGS; and 0.19 (0.02) and 0.21 for GSR.Features selected for the UGS model were: age, chair stands time, body mass index, grip strength, number of medications, resting state pulse interval, mean motor reaction time in the choice reaction time test, height, depression score, sit-to-stand difference in diastolic blood pressure, and left visual acuity.The features selected for the MGS model were: age, grip strength, repeated chair stands time, body mass index, education, mean motor reaction time in the choice reaction time test, number of medications, height, the standard deviation of the mean reaction time in the sustained attention to response task, mean heart rate at resting state, fear of falling, MOCA errors, orthostatic intolerance during active stand, smoking status, total heart beat power during paced breathing, the root mean square of successive differences between heartbeats during paced breathing, and visual acuity.Finally, the features chosen for the GSR model were: mean motor reaction time in the choice reaction time test, grip strength, education, chair stands time, MOCA errors, accuracy proportion in the sound induced flash illusion (two beeps and one flash with stimulus-onset asynchrony of +150 ms), fear of falling, height, age, sex, orthostatic intolerance, MMSE errors, and number of cardiovascular conditions.MGS and UGS were more explainable than GSR. All three models contain features from all five categories. There were common features to all three models (age, grip strength, chair stands time, mean motor reaction time in the choice reaction time test, and height), but also some features unique to each of them. Overall, findings on all three models were clinically plausible and support a network physiology approach to the understanding of predictors of performance-based tasks. By employing an explainable machine learning technique, our observations may help clinicians gain new insights into the multisystem predictors of gait speed and gait speed reserve in older adults.

Aim Fried's frailty phenotype (FP) is defined by exhaustion (EX), unexplained weight loss (WL), weakness (WK), slowness (SL) and low physical activity (LA). Three or more components define the frail state, and one or two the prefrail. We described longitudinal transitions of FP states and components in The Irish Longitudinal Study on Ageing (TILDA). Methods We included participants aged ≥50 years with FP information at TILDA wave 1 (2010), who were followed-up over four longitudinal waves (2012, 2014, 2016, 2018). Next-wave transition probabilities were estimated with multi-state Markov models. Results 5683 wave 1 participants were included (2612 men and 3071 women; mean age 63.1 years). Probabilities from non-frail to prefrail, and non-frail to frail were 27% and 2%, respectively. Prefrail had a 32% probability of reversal to non-frail, and a 10% risk of progression to frail. Frail had an 18% probability of reversal to prefrail and 31% risk of death. Probabilities of transitioning from not having to having a component were: 17% for LA, 11% for SL, 9% for EX, 7% for WL and 6% for WK. Probabilities of having a FP component and dying were: 17% for WL, 15% for WK, 14% for SL, 13% for EX, and 10% for LA. Probabilities of having a component and recovering at the next wave were: 59% for WL, 58% for EX, 40% for WK, 35% for LA and 23% for SL. Conclusions FP states and components are characterized by dynamic longitudinal transitions. Opportunities exist for reducing the probability of adverse transitions.

ObjectivesTo identify serum protein biomarkers which might separate early inflammatory arthritis (EIA) patients with psoriatic arthritis (PsA) from those with rheumatoid arthritis (RA) to provide an accurate diagnosis and support appropriate early intervention.MethodsIn an initial protein discovery phase, the serum proteome of a cohort of patients with PsA and RA was interrogated using unbiased liquid chromatography mass spectrometry (LC-MS/MS) (n=64 patients), a multiplexed antibody assay (Luminex) for 48 proteins (n=64 patients) and an aptamer-based assay (SOMAscan) targeting 1,129 proteins (n=36 patients). Subsequently, analytically validated targeted multiple reaction monitoring (MRM) assays were developed to further evaluate those proteins identified as discriminatory during the discovery. During an initial verification phase, MRM assays were developed to a panel of 150 proteins (by measuring a total of 233 peptides) and used to re-evaluate the discovery cohort (n=60). During a second verification phase, the panel of proteins was expanded to include an additional 23 proteins identified in other proteomic discovery analyses of arthritis patients. The expanded panel was evaluated using a second, independent cohort of PsA and RA patients (n=167).ResultsMultivariate analysis of the protein discovery data revealed that it was possible to discriminate PsA from RA patients with an area under the curve (AUC) of 0.94 for nLC-MS/MS, 0.69 for Luminex based measurements; 0.73 for SOMAscan analysis. During the initial verification phase, random forest models confirmed that proteins measured by MRM could differentiate PsA and RA patients with an AUC of 0.79 and during the second phase of verification the expanded panel could segregate the two disease groups with an AUC of 0.85.ConclusionWe report a serum protein biomarker panel which can separate EIA patients with PsA from those with RA. We suggest that the routine use of such a panel in EIA patients will improve clinical decision making and with continued evaluation and refinement using additional patient cohorts will support the development of a diagnostic test for patients with PsA.

The aging process is characterized by the presence of high interindividual variation between individuals of the same chronical age prompting a search for biomarkers that capture this heterogeneity. The present study examines the associations of four epigenetic clocks - Horvath, Hannum, PhenoAge, GrimAge - with a wide range of clinical phenotypes, and with all-cause mortality at up to 10-year follow-up in a sample of 490 participants in the Irish Longitudinal Study on Ageing. Results indicate that the GrimAge clock represents a step-improvement in the predictive utility of the epigenetic clocks for identifying age-related decline in an array of clinical phenotypes.

BackgroundOrthostatic hypotension (OH) can be assessed with non-invasive continuous beat-to-beat haemodynamic monitoring during active stand (AS) testing; this yields large volumes of data outside the scope of the traditional OH definition. We explored clinical associations of different AS patterns in participants from Wave 1 of the Irish Longitudinal Study on Ageing.MethodsAS patterns were generated based on three sequential binary systolic blood pressure features: drop ≥40 mmHg within 10 sec post-stand (“immediate deficit”), failure to return to within 20 mmHg of supine level at 40 sec after standing (“stabilisation deficit”) and drop ≥20 mmHg between >40 and 120 sec post-stand (“late deficit”). Eight AS groups resulted from combining the presence/absence of these three features. The groups were cross-sectionally characterised, and their ability to independently predict orthostatic intolerance (OI) during AS, and falls or syncope in the past year, was evaluated using multivariate logistic regression models.ResultsA total of 4,899 participants were included (mean age 61), of which 3,312 (68%) had no deficits. Older age was associated with stabilisation deficit and late deficits were seen in groups with higher proportions of beta blockers and psychotropic medications. Regression models identified independent associations between OI and three immediate-deficit groups; associations seemed stronger as more deficits were present. There was a significant association between falls history and the three-deficit group (odds ratio 1.54, 95% confidence interval: 1.15–2.07, P = 0.004).ConclusionsMore deficits seemed associated with the higher risk of OI and falls history. Observations are not causal but the recognition of these patterns may help clinicians focus on careful prescribing.

The world’s ageing population is rapidly increasing but people’s healthspan is not being sustained. Activities of daily living and Montreal Cognitive Assessment scores from the first wave of a large nationally representative longitudinal study in ageing (TILDA) were analysed using multiple correspondence analysis, k-means clustering, network analysis and association rules mining, to find latent patterns in the data and categorise disability among older adults. It was observed that 6.2% of the population had a greater degree of frailty, specifically cognitive impairment. Additionally, the overall population showed difficulty in performing physically demanding activities. Thus, self-reported ADLs have a diagnostic importance as they indicate the level of cognitive and physical functional decline in the older population.

Although there is some evidence suggesting that audiovisual integration is inefficient in older adults, and that such inefficiency is associated with age-related functions such as mild cognitive impairment, falls, and balance maintenance, these associations have yet to be demonstrated in a population-representative study of ageing. Based on a sample of 3,955 adults aged over 50 years, we investigated the role of age, cognitive status, and sex on susceptibility to the sound-induced flash illusion (SIFI) as a measure of audiovisual temporal integration, while controlling for a range of covariates. We developed a hierarchical Bayesian, ordinal-regression model to determine which variables predicted audiovisual integration. Higher susceptibility to the SIFI was predicted by older age, female sex (at larger temporal asynchronies), and a lower score on the Montreal Cognitive Assessment (MoCA). Our results confirm, in a population-representative sample, that enhanced audiovisual integration is associated with ageing and extend the association between multisensory integration and mild cognitive impairment to global cognitive status. Importantly, the findings also highlight the role of the sex of the participant as a previously overlooked factor in studying multisensory perception in ageing. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

ObjectivesThe International League of Associations for Rheumatology classification criteria define systemic juvenile idiopathic arthritis (SJIA) by the presence of fever, rash and chronic arthritis. Recent initiatives to revise current criteria recognise that a lack of arthritis complicates making the diagnosis early, while later a subgroup of patients develops aggressive joint disease. The proposed biphasic model of SJIA also implies a ‘window of opportunity’ to abrogate the development of chronic arthritis. We aimed to identify novel SJIA biomarkers during different disease phases.MethodsChildren with active SJIA were subgrouped clinically as systemic autoinflammatory disease with fever (SJIAsyst) or polyarticular disease (SJIApoly). A discovery cohort of n=10 patients per SJIA group, plus n=10 with infection, was subjected to unbiased label-free liquid chromatography mass spectrometry (LC-MS/MS) and immunoassay screens. In a separate verification cohort (SJIAsyst, n=45; SJIApoly, n=29; infection, n=32), candidate biomarkers were measured by multiple reaction monitoring MS (MRM-MS) and targeted immunoassays.ResultsSignatures differentiating the two phenotypes of SJIA could be identified. LC-MS/MS in the discovery cohort differentiated SJIAsyst from SJIApoly well, but less effectively from infection. Targeted MRM verified the discovery data and, combined with targeted immunoassays, correctly identified 91% (SJIAsyst vs SJIApoly) and 77% (SJIAsyst vs infection) of all cases.ConclusionsMolecular signatures differentiating two phenotypes of SJIA were identified suggesting shifts in underlying immunological processes in this biphasic disease. Biomarker signatures separating SJIA in its initial autoinflammatory phase from the main differential diagnosis (ie, infection) could aid early-stage diagnostic decisions, while markers of a phenotype switch could inform treat-to-target strategies.

Background The BIOmarkers of COMorbidities (BIOCOM) in Psoriasis (Pso) study is a longitudinal study in which we aim to identify clinical, genetic and protein biomarker features associated with the development of co-morbidities, notably psoriatic arthritis (PsA), in patients with psoriasis (Pso). Pso usually precedes the development of PsA with an average interval of 10 years. Thus, patients with Pso are an ideal group in which to study the early events in the evolution to PsA. Objectives To use a targeted proteomics approach to identify a serum protein, or panel of proteins, which can predict the development of PsA in patients with Pso. As a first step, we initially sought to identify serum proteins capable of discriminating between patients with Pso only and patients with established PsA. Methods 30 patients with Pso and 30 patients with established PsA were selected from the BIOCOM-Pso database. Serum samples from these patients, taken at their initial assessment, were digested using an established in-house standard operating protocol (SOP). Once digested, a targeted proteomics approach using liquid chromatography – mass spectrometry (LC-MS) and a multiple reaction monitoring (MRM) assay called PAPRICA was used to measure candidate biomarker proteins. These 206 proteins (423 peptides) were previously identified as being potential biomarkers in a number of different inflammatory rheumatological conditions. Results The demographics of the 2 patient groups are shown in Table 1. Targeted proteomics data from the PsA and Pso patients was subjected to univariate and multivariate analysis using a leave one out cross validation approach. The initial results revealed that the application of the PAPRICA method to the BIOCOM-Pso samples resulted in a dataset in which 88 of the 206 PAPRICA proteins could be reliably measured (CV Area 5; Library Dot Product > 0.8). This subset of the PAPRICA proteins was not able to discriminate between PsA and Pso and none of the associated peptides were significantly different between the two groups (p value Conclusion Analysis of a subset of 88 of the 206 biomarker proteins in the PAPRICA method, in patients with PsA and Pso, did not reveal peptides (proteins) that were statistically different between these two groups. Multivariate analysis generated a model that was unable to discriminate between patients with PsA and Pso. The possibility that the full PAPRICA assay may be able to discriminate between PsA and Pso will be explored, as will supplementing the PAPRICA method with additional biomarkers including proteins that may be identified in an unbiased proteome wide screen of PsA vs Pso serum samples. Disclosure of Interests None declared

Brief introduction Almost half of all deaths in Europe are attributable to by Cardiovascular Disease (CVD). There is a critical unmet need for better biomarkers so that CVD can be diagnosed at an earlier stage and with greater accuracy. Mass spectrometry-based multiple reaction monitoring (MRM) allows for rapid, targeted measurement of multiple protein biomarkers via detection of ‘proteotypic’ peptide fragments. Such assays can be developed in blood samples, which is desirable for minimally-invasive, routine monitoring of cardiovascular health. The purpose of this study was to design a robust MRM-based assay for the simultaneous detection and measurement of a panel of 36 proteins, including 19 known and 17 novel protein biomarkers of CVD, in blood. The overall aim was to evaluate the potential clinical utility of this biomarker panel for prediction of heart failure in a cohort of 500 patients. Explanation of basic methods Proteotypic peptides for each of the 36 candidate biomarker proteins were identified using Skyline and Spectrum Mill PeptideSelector software. MRM assays were developed using nanoflow reverse phase C18 chromatographic ChipCube based separation, coupled to a 6460 triple quadrupole mass spectrometer. Patient samples and synthetic peptides were used to obtain reference MS/MS spectra for building high quality MRM assays. For initial verification of the candidate protein biomarker panel, the MRM assay was applied, in a sample blinded manner, to a cohort of 500 serum samples from patients with heart failure and age-matched non-heart failure controls with CVD risk-factors. Results MRM assays were established for 35 peptides, corresponding to 9 of the known candidate biomarker proteins and all 16 of the novel biomarkers (Table 1). Individually, a number of the biomarker proteins show differential expression between heart failure with preserved and reduced ejection fraction (Table 2). Combined measurement of all biomarker proteins was found to have greater predictive capacity for heart failure than the current gold standard of B-type natriuretic peptide (BNP) alone. A statistical mode found that the biomarker panel, combined with BNP was capable of correctly predicting heart failure in blinded patient samples with accuracy of 82.19% AUC of 0.89 (Figure 1). Conclusions/Implications Through this study, assays have been developed for robust, multiplexed measurement of 26 CVD biomarker proteins in patient serum samples. Here it was observed that measurement of the full panel of CVD protein biomarkers offers more accurate prediction of heart failure than measurement of BNP alone. This panel will need to be further validated in independent patient cohorts. Conflict of interest None

Durability traits in Thoroughbred horses are heritable, economically valuable and may affect horse welfare. The aims of this study were to test the hypotheses that (i) durability traits are heritable and (ii) genetic data may be used to predict a horse's potential to have a racecourse start. Heritability for the phenotype ‘number of 2‐ and 3‐year‐old starts’ was estimated to be urn:x-wiley:02689146:media:age12798:age12798-math-0001 = 0.11 ± 0.02 (n = 4499). A genome‐wide association study identified SNP contributions to the trait. The neurotrimin (NTM), opioid‐binding protein/cell adhesion molecule like (OPCML) and prolylcarboxypeptidase (PRCP) genes were identified as candidate genes associated with the trait. NTM functions in brain development and has been shown to have been selected during the domestication of the horse. PRCP is an established expression quantitative trait locus involved in the interaction between voluntary exercise and body composition in mice. We hypothesise that variation at these loci contributes to the motivation of the horse to exercise, which may influence its response to the demands of the training and racing environment. A random forest with mixed effects (RFME) model identified a set of SNPs that contributed to 24.7% of the heritable variation in the trait. In an independent validation set (n = 528 horses), the cohort with high genetic potential for a racecourse start had significantly fewer unraced horses (16% unraced) than did low (27% unraced) potential horses and had more favourable race outcomes among those that raced. Therefore, the information from SNPs included in the model may be used to predict horses with a greater chance of a racecourse start.

Background Biomechanical screening assessments are used to provide useful information about an athlete's movement proficiency. Clinically, movement proficiency is typically evaluated visually. This can result in low levels of agreement, leading to difficulties in ensuring consistent athlete assessment. Objective The objective was to determine levels of agreement within and between physical therapists and physical therapist students when visually evaluating athletes’ movement proficiency during biomechanical screening assessments. Design This was an observational study. Methods Twenty-seven physical therapists and 20 physical therapist students assessed 100 video recordings of athletes performing 4 lower-extremity biomechanical screening assessments: squat, lunge, single leg squat, and deadlift. Analysis was completed on conditioned and unconditioned data. In the conditioned data, technique deviations were induced purposefully by the athletes. In the unconditioned data, deviations occurred naturally due to increased weight or movement complexity. In order to determine levels of agreement in the assessments, participants were required to classify the athletes’ movement as acceptable or aberrant. Each participant assessed the same video recordings on 2 separate occasions at least 30 days apart. Agreement levels were determined using Cohen κ and Fleiss κ. Results Kappa scores at an interrater level ranged from 0.18 to 0.53, and intrarater agreement ranged from 0.38 to 0.62. Levels of agreement were higher in the conditioned data compared with the unconditioned data. Overall, the lunge and squat produced higher levels of agreement than the deadlift and single-leg squat. Students and physical therapists demonstrated similar levels of agreement. Limitations Screening assessments were evaluated through the use of video analysis. Conclusions Greater efforts are needed to ensure standardization of movement analysis.

Background Systemic juvenile idiopathic arthritis (SJIA) is a childhood rheumatic auto-inflammatory disorder of largely unknown pathogenesis. The presence of fever, rash and arthritis support a diagnosis of SJIA, though early in disease arthritis may be minimal, complicating the exclusion of alternative diagnoses such as infection. Furthermore, two clinical phenotypes of SJIA can be identified, a chronic articular-dominant (ART_SJIA) and a classical auto-inflammatory phenotype (AID_SJIA). Objectives To identify novel serum protein biomarkers that may discriminate ART_SJIA from AID_SJIA and distinguish AID_SJIA from infection. Methods Patients with active SJIA (joint activity plus or minus fever and elevated laboratory inflammation markers) were sub-grouped into the two clinical phenotypes (ART_SJIA and AID_SJIA). Serum from patients with SJIA or confirmed infection was analysed for the standard laboratory markers: C-reactive protein (CRP), white cell count (WCC) and erythrocyte sedimentation rate (ESR). A “discovery cohort” (n=10 per group) of patient serum samples was subjected to unbiased label-free proteomics using liquid chromatography mass spectrometry (LC-MS/MS) and in a separate “verification cohort” (AID_SJIA, n=48; ART_SJIA, n=29; infection, n=32) candidate biomarkers were measured by multiple reaction monitoring MS (MRM-MS; Agilent 6490 and 6495) and microsphere bead-based immunoassay (Luminex). Serum concentrations of S100A12 and MRP8/14 were also measured in all samples using enzyme linked immunoabsorbant assays (ELISAs). Results The routine laboratory markers CRP, WCC, ESR, as well as ELISA-measured S100A12 and MRP8/14 serum concentrations were highest in AID_SJIA, followed by infection and were lowest in patients with ART_SJIA. Proteins identified and quantified by LC-MS/MS could differentiate patients with ART_SJIA from those with AID_SJIA (area under the curve, AUC: 0.97). The discrimination between AID_SJIA and infection performed less well, AUC: 0.58. Targeted MRM measurement of novel protein candidate biomarkers verified the discovery cohort data. A combined biomarker panel consisting of MRM, ELISA and Luminex analysis which was evaluated using a Random forest model, indicated the correct overall identification of the clinical groups to be as follows: ART_SJIA: 24/29 (83%), AID_SJIA: 37/45 (82%) and infection: 20/32 (63%). Conclusions Significant differences in the serum protein signature between two phenotypes of SJIA suggest that different immunological processes may underlie these phenotypes. Serum protein profiles also distinguished patients with SJIA and infection. Distinguishing these two groups remains an important goal of research and therefore this study could help inform future prospective studies. Disclosure of Interest None declared

Increasing experimental and clinical evidence suggest a contribution of non-drug related risk factors (e.g., underlying disease, bacterial/viral infection) to idiosyncratic drug reactions (IDR). Our previous work showed that co-treatment with bacterial endotoxin (LPS) and therapeutic doses of diclofenac (Dcl), an analgesic associated with drug idiosyncrasy in patients, induced severe hepatotoxicity in rats. Here, we used an integrated discovery to targeted LC–MS proteomics approach to identify mechanistically relevant liver and plasma proteins modulated by LPS/Dcl treatment, potentially applicable as early markers for IDRs. Based on pre-screening results and their role in liver toxicity, 47 liver and 15 plasma proteins were selected for targeted LC–MS analysis. LPS alone significantly changed the levels of 19 and 3 of these proteins, respectively. T-kininogen-1, previously suggested as a marker of drug-induced liver injury, was markedly elevated in plasma after repeated Dcl treatment in the absence of hepatotoxicity, possibly indicating clinically silent stress. Dcl both alone and in combination with LPS, caused up-regulation of the ATP synthase subunits (ATP5J, ATPA, and ATPB), suggesting that Dcl may sensitize cells against additional stress factors, such as LPS through generation of mitochondrial stress. Additionally, depletion of plasma fibrinogen was observed in the co-treatment group, consistent with an increased hepatic fibrin deposition and suspected contribution of the hemostatic system to IDRs. In contrast, several proteins previously suggested as liver biomarkers, such as clusterin, did not correlate with liver injury in this model. Taken together, these analyses revealed proteomic changes in a rat model of LPS/Dcl co-administration that could offer mechanistic insight and may serve as biomarkers or safety alert for a drug’s potential to cause IDRs.

Background Psoriatic Arthritis (PsA) is an inflammatory arthritis (IA) frequently associated with psoriasis. Clinically, it is a complex heterogeneous disease and there are no diagnostic tests or criteria [1]. At first presentation, PsA may resemble other disease types - especially rheumatoid arthritis (RA). Making an accurate and early diagnosis is particularly important to ensure that individual patients receive effective and safe medication and so optimise long-term patient outcomes. Thus, it is widely acknowledged by physicians and patients alike that a new diagnostic test is needed to facilitate the early and specific diagnosis of PsA [2]. Objectives To (i) identify and verify candidate biomarkers with the potential to segregate patients with PsA from those with RA; and (ii) explore the value of combining different proteomic discovery platforms. Methods Serum samples were obtained from a cohort of 64 patients (32 PsA and 32 RA) defined as early onset ( Results In this study, it was possible to quantify 387, 48 and 1129 proteins from LC-MS/MS, Luminex and SOMAscan analysis, respectively. Proteins with the ability to segregate PsA patients from those with RA were identified by random forest analysis; LC-MS/MS (AUC 0.94), Luminex (AUC 0.69) and SOMAscan (AUC 0.73). The application of the random forest model to the (i) combined data and (ii) MRM data set is part of ongoing work. Conclusions To date, statistical analysis revealed LC-MS/MS identified proteins were the most discriminatory. An MRM assay has been developed to the top 54 LC-MS/MS proteins and this assay has been applied to the discovery cohort (data analysis ongoing). The assay will next be applied to additional evaluation cohorts that include patients with spondyloarthritis and psoriasis. Discriminatory proteins verified here represent candidates for inclusion in a blood based multi-analyte test that could ultimately be used in the diagnosis of PsA. References Butt AQ, Mc Ardle A, Gibson DS, FitzGerald OF, Pennington SR. Psoriatic arthritis under a proteomic spotlight: application of novel technologies to advance diagnosis and management. Curr Rheumatology Reports 2015;17(5):35. doi: 1007/ss11926-015-0509–0. Mc Ardle, A. Butt AQ, Szentpetery, A. De Jager, Wilco. De Roock, Sytze. FitzGerald, O. Pennington, SR. Developing Clinically Relevant Biomarkers in Inflammatory Arthritis: A Multiplatform Approach for Serum Candidate Protein Discovery. Proteomics Clinical Applications, 2015. doi: 10.1002/prca.201500046. Disclosure of Interest None declared

Objective Biological therapies, which include antitumour necrosis factor-α and T-cell inhibitors, are potentially effective treatments for psoriatic arthritis (PsA) but are costly and may induce a number of side effects. Response to treatment in PsA is variable and difficult to predict. Here, we sought to identify a panel of protein biomarkers that could be used to predict which patients diagnosed with PsA will respond to biologic treatment. Methods An integrated discovery to targeted proteomics approach was used to investigate the protein profiles of good and non-responders to biological treatments in patients with PsA. Reverse-phase liquid chromatography coupled to tandem mass spectrometry was used to generate protein profiles of synovial tissue obtained at baseline from 10 patients with PsA. Targeted proteomics using multiple reaction monitoring (MRM) was used to confirm and prevalidate a potential protein biomarker panel in 18 and 7 PsA patient samples, respectively. Results A panel of 107 proteins was selected, and targeted mass spectrometry MRM assays were successfully developed for 57 of the proteins. The 57 proteins include S100-A8, S100-A10, Ig kappa chain C fibrinogen-α and γ, haptoglobin, annexin A1 and A2, collagen alpha-2, vitronectin, and alpha-1 acid glycoprotein. The proteins were measured simultaneously and confirmed to be predictive of response to treatment with an area under the curve of 0.76. In a blinded study using a separate cohort of patients, the panel was able to predict response to treatment. Conclusions The approach reported here and the initial data provide evidence that a multiplexed protein assay of a panel of biomarkers that predict response to treatment could be developed. Trial registration number ISRCTN23328456.

Background Identification of gene expression profiles that differentiate experimental groups is critical for discovery and analysis of key molecular pathways and also for selection of robust diagnostic or prognostic biomarkers. While integration of differential expression statistics has been used to refine gene set enrichment analyses, such approaches are typically limited to single gene lists resulting from simple two-group comparisons or time-series analyses. In contrast, functional class scoring and machine learning approaches provide powerful alternative methods to leverage molecular measurements for pathway analyses, and to compare continuous and multi-level categorical factors. Results We introduce GOexpress, a software package for scoring and summarising the capacity of gene ontology features to simultaneously classify samples from multiple experimental groups. GOexpress integrates normalised gene expression data (e.g., from microarray and RNA-seq experiments) and phenotypic information of individual samples with gene ontology annotations to derive a ranking of genes and gene ontology terms using a supervised learning approach. The default random forest algorithm allows interactions between all experimental factors, and competitive scoring of expressed genes to evaluate their relative importance in classifying predefined groups of samples. Conclusions GOexpress enables rapid identification and visualisation of ontology-related gene panels that robustly classify groups of samples and supports both categorical (e.g., infection status, treatment) and continuous (e.g., time-series, drug concentrations) experimental factors. The use of standard Bioconductor extension packages and publicly available gene ontology annotations facilitates straightforward integration of GOexpress within existing computational biology pipelines. Electronic supplementary material The online version of this article (doi:10.1186/s12859-016-0971-3) contains supplementary material, which is available to authorized users.

Bayesian Additive Regression Trees (BART) is a statistical sum of trees model. It can be considered a Bayesian version of machine learning tree ensemble methods where the individual trees are the base learners. However for data sets where the number of variables $p$ is large (e.g. $p>5,000$) the algorithm can become prohibitively expensive, computationally. Another method which is popular for high dimensional data is random forests, a machine learning algorithm which grows trees using a greedy search for the best split points. However, as it is not a statistical model, it cannot produce probabilistic estimates or predictions. We propose an alternative algorithm for BART called BART-BMA, which uses Bayesian Model Averaging and a greedy search algorithm to produce a model which is much more efficient than BART for datasets with large $p$. BART-BMA incorporates elements of both BART and random forests to offer a model-based algorithm which can deal with high-dimensional data. We have found that BART-BMA can be run in a reasonable time on a standard laptop for the "small $n$ large $p$" scenario which is common in many areas of bioinformatics. We showcase this method using simulated data and data from two real proteomic experiments; one to distinguish between patients with cardiovascular disease and controls and another to classify agressive from non-agressive prostate cancer. We compare our results to their main competitors. Open source code written in R and Rcpp to run BART-BMA can be found at: https://github.com/BelindaHernandez/BART-BMA.git

Proteomic biomarker discovery has led to the identification of numerous potential candidates for disease diagnosis, prognosis, and prediction of response to therapy. However, very few of these identified candidate biomarkers reach clinical validation and go on to be routinely used in clinical practice. One particular issue with biomarker discovery is the identification of significantly changing proteins in the initial discovery experiment that do not validate when subsequently tested on separate patient sample cohorts. Here, we seek to highlight some of the statistical challenges surrounding the analysis of LC-MS proteomic data for biomarker candidate discovery. We show that common statistical algorithms run on data with low sample sizes can overfit and yield misleading misclassification rates and AUC values. A common solution to this problem is to prefilter variables (via, e.g. ANOVA and or use of correction methods such as Bonferonni or false discovery rate) to give a smaller dataset and reduce the size of the apparent statistical challenge. However, we show that this exacerbates the problem yielding even higher performance metrics while reducing the predictive accuracy of the biomarker panel. To illustrate some of these limitations, we have run simulation analyses with known biomarkers. For our chosen algorithm (random forests), we show that the above problems are substantially reduced if a sufficient number of samples are analyzed and the data are not prefiltered. Our view is that LC-MS proteomic biomarker discovery data should be analyzed without prefiltering and that increasing the sample size in biomarker discovery experiments should be a very high priority.

Background Psoriatic Arthritis (PsA) is a form of seronegative inflammatory arthritis (IA) frequently associated with psoriasis. Similar to other forms of IA, it is a complex heterogeneous disease. While PsA can be characterised by distinct clinical manifestations, at early onset, it often resembles other disease types - especially rheumatoid arthritis (RA). During the treatment and management of the disease important clinical decisions are made and these have a significant impact on patient outcomes. Making an accurate and early diagnosis is particularly important to ensuring that individual patients receive effective and safe medication. Thus, it is widely acknowledged by physicians and patients alike that a new test is needed to facilitate the early and specific diagnosis of PsA [1]. Objectives To (i) identify candidate biomarkers with the potential to segregate patients with PsA from those with RA (ii) explore the value of combining different omic discovery platforms. Methods Serum samples were obtained from a cohort of 64 patients (32 PsA and 32 RA) defined as early onset ( Results In this study, by combining 4 different omic technologies it was possible to quantify 376, 48, 1129, 387 analytes by miRNA, Luminex, SOMAscan and LC-MS/MS analysis, respectively. Statistical analysis revealed that a total of 287 analytes were differentially expressed across the two conditions (p≤0.05). Furthermore, random forest analysis revealed the most discriminatory biomarkers from the miRNA, Luminex and SOMAscan analysis. Figure 1 A-C illustrates the receiver operating characteristic (ROC) curves generated from these analyses. The application of the random forest model to the LC-MS/MS data is part of ongoing work. Conclusions By combining the most discriminatory analytes from each type of analysis into a single algorithm, a biomarker signature with increased predictive potency should be identified. Thus, these molecules represent candidates for inclusion into blood based mulit-analyte test that could be used in the diagnosis of PsA. References Mc Ardle. A. Butt, AQ. Szentpetery, De Jager, Wilco. De Rook, Sytze. FitzGerald, O. Pennington, SR. Developing Clincially Relevant Biomarerks In Inflammatory Arthritis: A Mutli-Platform Approach for Serum Candidate Protein Discovery. Clincal Proteomics 2015, doi: 10.1oo2/prca.201500046 Acknowledgement Kieran Wynne is thanked for the technical support provided during this project. Disclosure of Interest None declared