Disease progression modelling of Alzheimer's disease using probabilistic principal components analysis.

The recent biological redefinition of Alzheimer's Disease (AD) has spurred the development of statistical models that relate changes in biomarkers with neurodegeneration and worsening condition linked to AD. The ability to measure such changes may facilitate earlier diagnoses for affected individuals and help in monitoring the evolution of their condition. Amongst such statistical tools, disease progression models (DPMs) are quantitative, data-driven methods that specifically attempt to describe the temporal dynamics of biomarkers relevant to AD. Due to the heterogeneous nature of this disease, with patients of similar age experiencing different AD-related changes, a challenge facing longitudinal mixed-effects-based DPMs is the estimation of patient-realigning time-shifts. These time-shifts are indispensable for meaningful biomarker modelling, but may impact fitting time or vary with missing data in jointly estimated models. In this work, we estimate an individual's progression through Alzheimer's disease by combining multiple biomarkers into a single value using a probabilistic formulation of principal components analysis. Our results show that this variable, which summarises AD through observable biomarkers, is remarkably similar to jointly estimated time-shifts when we compute our scores for the baseline visit, on cross-sectional data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Reproducing the expected properties of clinical datasets, we confirm that estimated scores are robust to missing data or unavailable biomarkers. In addition to cross-sectional insights, we can model the latent variable as an individual progression score by repeating estimations at follow-up examinations and refining long-term estimates as more data is gathered, which would be ideal in a clinical setting. Finally, we verify that our score can be used as a pseudo-temporal scale instead of age to ignore some patient heterogeneity in cohort data and highlight the general trend in expected biomarker evolution in affected individuals.

An evaluation of existing text de-identification tools for use with patient progress notes from Australian general practice.

INTRODUCTION: Digitized patient progress notes from general practice represent a significant resource for clinical and public health research but cannot feasibly and ethically be used for these purposes without automated de-identification. Internationally, several open-source natural language processing tools have been developed, however, given wide variations in clinical documentation practices, these cannot be utilized without appropriate review. We evaluated the performance of four de-identification tools and assessed their suitability for customization to Australian general practice progress notes. METHODS: Four tools were selected: three rule-based (HMS Scrubber, MIT De-id, Philter) and one machine learning (MIST). 300 patient progress notes from three general practice clinics were manually annotated with personally identifying information. We conducted a pairwise comparison between the manual annotations and patient identifiers automatically detected by each tool, measuring recall (sensitivity), precision (positive predictive value), f1-score (harmonic mean of precision and recall), and f2-score (weighs recall 2x higher than precision). Error analysis was also conducted to better understand each tool's structure and performance. RESULTS: Manual annotation detected 701 identifiers in seven categories. The rule-based tools detected identifiers in six categories and MIST in three. Philter achieved the highest aggregate recall (67%) and the highest recall for NAME (87%). HMS Scrubber achieved the highest recall for DATE (94%) and all tools performed poorly on LOCATION. MIST achieved the highest precision for NAME and DATE while also achieving similar recall to the rule-based tools for DATE and highest recall for LOCATION. Philter had the lowest aggregate precision (37%), however preliminary adjustments of its rules and dictionaries showed a substantial reduction in false positives. CONCLUSION: Existing off-the-shelf solutions for automated de-identification of clinical text are not immediately suitable for our context without modification. Philter is the most promising candidate due to its high recall and flexibility however will require extensive revising of its pattern matching rules and dictionaries.

Identification of Leukocyte Surface P2X7 as a Biomarker Associated with Alzheimer's Disease.

Alzheimer's disease (AD) has shown altered immune responses in the periphery. We studied P2X7 (a proinflammatory receptor and a scavenger receptor) and two integrins, CD11b and CD11c, on the surface of circulating leukocytes and analysed their associations with Aß-PET, brain atrophy, neuropsychological assessments, and cerebrospinal fluid (CSF) biomarkers. Total 287 age-matched, sex-balanced participants were recruited in a discovery cohort and two validation cohorts through the AIBL study and studied using tri-colour flow cytometry. Our results demonstrated reduced expressions of P2X7, CD11b, and CD11c on leukocytes, particularly monocytes, in Aß +ve cases compared with Aß -ve controls. P2X7 and integrin downregulation was observed at pre-clinical stage of AD and stayed low throughout disease course. We further constructed a polygenic risk score (PRS) model based on 12 P2RX7 risk alleles to assess the genetic impact on P2X7 function in AIBL and ADNI cohorts. No significant association was identified between the P2RX7 gene and AD, indicating that P2X7 downregulation in AD is likely caused by environmental changes rather than genetic factors. In conclusion, the downregulation of P2X7 and integrins at pre-clinical stage of AD indicates altered pro-inflammatory responses, phagocytic functions, and migrating capabilities of circulating monocytes in early AD pathogenesis. Our study not only improves our understanding of peripheral immune involvement in early stage of AD but also provides more insights into novel biomarker development, diagnosis, and prognosis of AD.

Process mining-driven analysis of COVID-19's impact on vaccination patterns.

Process mining is a discipline sitting between data mining and process science, whose goal is to provide theoretical methods and software tools to analyse process execution data, known as event logs. Although process mining was originally conceived to facilitate business process management activities, research studies have shown the benefit of leveraging process mining in healthcare contexts. However, applying process mining tools to analyse healthcare process execution data is not straightforward. In this paper, we show a methodology to: i) prepare general practice healthcare process data for conducting a process mining analysis; ii) select and apply suitable process mining solutions for successfully executing the analysis; and iii) extract valuable insights from the obtained results, alongside leads for traditional data mining analysis. By doing so, we identified two major challenges when using process mining solutions for analysing healthcare process data, and highlighted benefits and limitations of the state-of-the-art process mining techniques when dealing with highly variable processes and large data-sets. While we provide solutions to the identified challenges, the overarching goal of this study was to detect differences between the patients' health services utilization pattern observed in 2020-during the COVID-19 pandemic and mandatory lock-downs -and the one observed in the prior four years, 2016 to 2019. By using a combination of process mining techniques and traditional data mining, we were able to demonstrate that vaccinations in Victoria did not drop drastically-as other interactions did. On the contrary, we observed a surge of influenza and pneumococcus vaccinations in 2020, as opposed to other research findings of similar studies conducted in different geographical areas.

"Escalibur"-A practical pipeline for the de novo analysis of nucleotide variation in nonmodel eukaryotes.

The revolution in genomics has enabled large-scale population genetic investigations of a wide range of organisms, but there has been a relatively limited focus on improving analytical pipelines. To efficiently analyse large data sets, highly integrated and automated software pipelines, which are easy to use, efficient, reliable, reproducible and run in multiple computational environments, are required. A number of software workflows have been developed to handle and process such data sets for population genetic analyses, but effective, specialized pipelines for genetic and statistical analyses of nonmodel organisms are lacking. For most species, resources for variomes (sets of genetic variations found in populations of species) are not available, and/or genome assemblies are often incomplete and fragmented, complicating the selection of the most suitable reference genome when multiple assemblies are available. Additionally, the biological samples used often contain extraneous DNA from sources other than the species under investigation (e.g., microbial contamination), which needs to be removed prior to genetic analyses. For these reasons, we established a new pipeline, called Escalibur, which includes: functionalities, such as data trimming and mapping; selection of a suitable reference genome; removal of contaminating read data; recalibration of base calls; and variant-calling. Escalibur uses a proven gatk variant caller and workflow description language (WDL), and is, therefore, a highly efficient and scalable pipeline for the genome-wide identification of nucleotide variation in eukaryotes. This pipeline is available at https://gitlab.unimelb.edu.au/bioscience/escalibur (version 0.3-beta) and is essentially applicable to any prokaryote or eukaryote.

A blood-based signature of cerebrospinal fluid Aß1-42 status.

It is increasingly recognized that Alzheimer's disease (AD) exists before dementia is present and that shifts in amyloid beta occur long before clinical symptoms can be detected. Early detection of these molecular changes is a key aspect for the success of interventions aimed at slowing down rates of cognitive decline. Recent evidence indicates that of the two established methods for measuring amyloid, a decrease in cerebrospinal fluid (CSF) amyloid ß1-42 (Aß1-42) may be an earlier indicator of Alzheimer's disease risk than measures of amyloid obtained from Positron Emission Tomography (PET). However, CSF collection is highly invasive and expensive. In contrast, blood collection is routinely performed, minimally invasive and cheap. In this work, we develop a blood-based signature that can provide a cheap and minimally invasive estimation of an individual's CSF amyloid status using a machine learning approach. We show that a Random Forest model derived from plasma analytes can accurately predict subjects as having abnormal (low) CSF Aß1-42 levels indicative of AD risk (0.84 AUC, 0.78 sensitivity, and 0.73 specificity). Refinement of the modeling indicates that only APOEε4 carrier status and four plasma analytes (CGA, Aß1-42, Eotaxin 3, APOE) are required to achieve a high level of accuracy. Furthermore, we show across an independent validation cohort that individuals with predicted abnormal CSF Aß1-42 levels transitioned to an AD diagnosis over 120 months significantly faster than those with predicted normal CSF Aß1-42 levels and that the resulting model also validates reasonably across PET Aß1-42 status (0.78 AUC). This is the first study to show that a machine learning approach, using plasma protein levels, age and APOEε4 carrier status, is able to predict CSF Aß1-42 status, the earliest risk indicator for AD, with high accuracy.

A hybrid approach for automated mutation annotation of the extended human mutation landscape in scientific literature.

As the cost of DNA sequencing continues to fall, an increasing amount of information on human genetic variation is being produced that could help progress precision medicine. However, information about such mutations is typically first made available in the scientific literature, and is then later manually curated into more standardized genomic databases. This curation process is expensive, time-consuming and many variants do not end up being fully curated, if at all. Detecting mutations in the literature is the first key step towards automating this process. However, most of the current methods have focused on identifying mutations that follow existing nomenclatures. In this work, we show that there is a large number of mutations that are missed by using this standard approach. Furthermore, we implement the first mutation annotator to cover an extended mutation landscape, and we show that its F1 performance is the same performance as human annotation (F1 78.29 for manual annotation vs F1 79.56 for automatic annotation).

CuII(atsm) improves the neurological phenotype and survival of SOD1G93A mice and selectively increases enzymatically active SOD1 in the spinal cord.

Ubiquitous expression of mutant Cu/Zn-superoxide dismutase (SOD1) selectively affects motor neurons in the central nervous system (CNS), causing the adult-onset degenerative disease amyotrophic lateral sclerosis (ALS). The CNS-specific impact of ubiquitous mutant SOD1 expression is recapitulated in transgenic mouse models of the disease. Here we present outcomes for the metallo-complex CuII(atsm) tested for therapeutic efficacy in mice expressing SOD1G93A on a mixed genetic background. Oral administration of CuII(atsm) delayed the onset of neurological symptoms, improved locomotive capacity and extended overall survival. Although the ALS-like phenotype of SOD1G93A mice is instigated by expression of the mutant SOD1, we show the improved phenotype of the CuII(atsm)-treated animals involves an increase in mature mutant SOD1 protein in the disease-affected spinal cord, where concomitant increases in copper and SOD1 activity are also evident. In contrast to these effects in the spinal cord, treating with CuII(atsm) had no effect in liver on either mutant SOD1 protein levels or its activity, indicating a CNS-selective SOD1 response to the drug. These data provide support for CuII(atsm) as a treatment option for ALS as well as insight to the CNS-selective effects of mutant SOD1.

Predicting Alzheimer disease from a blood-based biomarker profile: A 54-month follow-up.

OBJECTIVE: We assessed a blood-based signature, which previously demonstrated high accuracy at stratifying individuals with high or low neocortical ß-amyloid burden (NAB), to determine whether it could also identify individuals at risk of progression to Alzheimer disease (AD) within 54 months. METHODS: We generated the blood-based signature for 585 healthy controls (HCs) and 74 participants with mild cognitive impairment (MCI) from the Australian Imaging, Biomarkers and Lifestyle Study who underwent clinical reclassification (blinded to biomarker findings) at 54-month follow-up. The individuals were split into estimated high and low NAB groups based on a cutoff of 1.5 standardized uptake value ratio. We assessed the predictive accuracy of the high and low NAB groupings based on progression to mild cognitive impairment or AD according to clinical reclassification at 54-month follow-up. RESULTS: Twelve percent of HCs with estimated high NAB progressed in comparison to 5% of HCs with estimated low NAB (odds ratio = 2.4). Forty percent of the participants with MCI who had estimated high NAB progressed in comparison to 5% of the participants with MCI who had estimated low NAB (odds ratio = 12.3). These ratios are in line with those reported for Pittsburgh compound B-PET results. Individuals with estimated high NAB had faster rates of memory decline than those with estimated low NAB. CONCLUSION: These findings suggest that a simple blood-based signature not only provides estimates of NAB but also predicts cognitive decline and disease progression, identifying individuals at risk of progressing toward AD at the prodromal and preclinical stages.

Lead and manganese levels in serum and erythrocytes in Alzheimer's disease and mild cognitive impairment: results from the Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing.

We examined serum and erythrocyte lead and manganese levels in the Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing (AIBL), which contains over 1000 registrants including over 200 cases of Alzheimer's disease (AD) and 100 mildly cognitively impaired (MCI) individuals. After correcting for confounding effects of age, collection site and sex, we found a significant decrease in serum manganese levels in AD subjects compared to healthy controls. Analysis of smaller subset of erythrocytes revealed no difference in either lead or manganese levels in AD. Although lead and manganese have neurotoxic effects and may be involved in AD pathology, our results showed that neither metal in serum nor erythrocytes are suitable biomarkers in our cohort. However, prospective studies might reveal whether the burden of either metal modifies disease outcomes.

Ferritin levels in the cerebrospinal fluid predict Alzheimer's disease outcomes and are regulated by APOE.

Brain iron elevation is implicated in Alzheimer's disease (AD) pathogenesis, but the impact of iron on disease outcomes has not been previously explored in a longitudinal study. Ferritin is the major iron storage protein of the body; by using cerebrospinal fluid (CSF) levels of ferritin as an index, we explored whether brain iron status impacts longitudinal outcomes in the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort. We show that baseline CSF ferritin levels were negatively associated with cognitive performance over 7 years in 91 cognitively normal, 144 mild cognitive impairment (MCI) and 67 AD subjects, and predicted MCI conversion to AD. Ferritin was strongly associated with CSF apolipoprotein E levels and was elevated by the Alzheimer's risk allele, APOE-ɛ4. These findings reveal that elevated brain iron adversely impacts on AD progression, and introduce brain iron elevation as a possible mechanism for APOE-ɛ4 being the major genetic risk factor for AD.

Decreased plasma iron in Alzheimer's disease is due to transferrin desaturation.

Plasma iron levels are decreased in Alzheimer's disease (AD) and associated with an idiopathic anemia. We examined iron-binding plasma proteins from AD patients and healthy controls from the Australian Imaging, Biomarkers and Lifestyle (AIBL) Flagship Study of Ageing using size exclusion chromatography-inductively coupled plasma-mass spectrometry. Peak area corresponding to transferrin (Tf) saturation was directly compared to routine pathological testing. We found a significant decrease in transferrin-associated iron in AD that was missed by routine pathological tests of transferrin saturation, and that was able to discriminate between AD and controls. The AD cases showed no significant difference in transferrin concentration, only a decrease in total transferrin-bound iron. These findings support that a previously identified decrease in plasma iron levels in AD patients within the AIBL study is attributable to decreased loading of iron into transferrin, and that this subtle but discriminatory change is not observed through routine pathological testing.

Peripheral α-defensins 1 and 2 are elevated in Alzheimer's disease.

Biomarkers enabling the preclinical identification of Alzheimer's disease (AD) remain one of the major unmet challenges in the field. The blood cellular fractions offer a viable alternative to current cerebrospinal fluid and neuroimaging modalities. The current study aimed to replicate our earlier reports of altered binding within the AD-affected blood cellular fraction to copper-loaded immobilized metal affinity capture (IMAC) arrays. IMAC and anti-amyloid-ß (Aß) antibody arrays coupled with mass spectrometry were used to analyze blood samples collected from 218 participants from within the AIBL Study of Aging. Peripheral Aß was fragile and prone to degradation in the AIBL samples, even when stored at -80°C. IMAC analysis of the AIBL samples lead to the isolation and identification of alpha-defensins 1 and 2 at elevated levels in the AD periphery, validating earlier findings. Alpha-defensins 1 and 2 were elevated in AD patients indicating that an inflammatory phenotype is present in the AD periphery; however, peripheral Aß levels are required to supplement their prognostic power.

Altered transition metal homeostasis in Niemann-Pick disease, type C1.

The loss of NPC1 protein function is the predominant cause of Niemann-Pick type C1 disease (NP-C1), a systemic and neurodegenerative disorder characterized by late-endosomal/lysosomal accumulation of cholesterol and other lipids. Limited evidence from post-mortem human tissues, an Npc1(-/-) mouse model, and cell culture studies also suggest failure of metal homeostasis in NP-C1. To investigate these findings, we performed a comprehensive transition metal analysis of cerebrospinal fluid (CSF), plasma and tissue samples from human NP-C1 patients and an Npc1(-/-) mouse model. NPC1 deficiency in the Npc1(-/-) mouse model resulted in a perturbation of transition metal homeostasis in the plasma and key organs (brain, liver, spleen, heart, lungs, and kidneys). Analysis of human patient CSF, plasma and post-mortem brain tissues also indicated disrupted metal homeostasis. There was a disparity in the direction of metal changes between the human and the Npc1(-/-) mouse samples, which may reflect species-specific metal metabolism. Nevertheless, common to both species is brain zinc accumulation. Furthermore, treatment with the glucosylceramide synthase inhibitor miglustat, the only drug shown in a controlled clinical trial to have some efficacy for NP-C1, did not correct the alterations in CSF and plasma transition metal and ceruloplasmin (CP) metabolism in NP-C1 patients. These findings highlight the importance of NPC1 function in metal homeostasis, and indicate that metal-targeting therapy may be of value as a treatment for NP-C.

Changes in plasma amyloid beta in a longitudinal study of aging and Alzheimer's disease.

BACKGROUND: A practical biomarker is required to facilitate the preclinical diagnosis of Alzheimer's disease (AD). METHODS: Plasma amyloid beta (Aß)1-40, Aß1-42, Aßn-40, and Aßn-42 peptides were measured at baseline and after 18 months in 771 participants from the Australian Imaging Biomarkers and Lifestyle (AIBL) study of aging. Aß peptide levels were compared with clinical pathology, neuroimaging and neuropsychological measurements. RESULTS: Although inflammatory and renal function covariates influenced plasma Aß levels significantly, a decrease in Aß1-42/Aß1-40 was observed in patients with AD, and was also inversely correlated with neocortical amyloid burden. During the 18 months, plasma Aß1-42 decreased in subjects with mild cognitive impairment (MCI) and in those transitioning from healthy to MCI. CONCLUSION: Our findings are consistent with a number of published plasma Aß studies and, although the prognostic value of individual measures in any given subject is limited, the diagnostic contribution of plasma Aß may demonstrate utility when combined with a panel of peripheral biomarkers.

Among vitamin B12 deficient older people, high folate levels are associated with worse cognitive function: combined data from three cohorts.

BACKGROUND: Folate fortification of food aims to reduce the number of babies born with neural tube defects, but has been associated with cognitive impairment when vitamin B12 levels are deficient. Given the prevalence of low vitamin B12 levels among the elderly, and the global deployment of food fortification programs, investigation of the associations between cognitive impairment, vitamin B12, and folate are needed. OBJECTIVE: To investigate the associations of serum vitamin B12, red cell folate, and cognitive impairment. METHODS: Data were collected on 1,354 subjects in two studies investigating cognitive impairment, and from patients attending for assessment or management of memory problems in the Barwon region of south eastern Australia between 2001 and 2011. Eligible subjects who had blood measurements of vitamin B12 and red cell folate taken within six months of cognitive testing were included. Subjects with stroke or neurodegenerative diseases other than Alzheimer's disease were excluded. A Mini-Mental State Examination score of <24 was used to define impaired cognitive function. RESULTS: Participants with low serum vitamin B12 (<250 pmol/L) and high red cell folate (>1,594 nmol/L) levels were more likely to have impaired cognitive performance (adjusted odds ratio (AOR) 3.45, 95% confidence interval (CI): 1.60-7.43, p = 0.002) when compared to participants with biochemical measurements that were within the normal ranges. Participants with high folate levels, but normal serum vitamin B12, were also more likely to have impaired cognitive performance (AOR 1.74, 95% CI: 1.03-2.95, p = 0.04). CONCLUSIONS: High folate or folic acid supplements may be detrimental to cognition in older people with low vitamin B12 levels. This topic is of global significance due to the wide distribution of food fortification programs, so prospective studies should be a high priority.

Increased risk of cognitive impairment in patients with diabetes is associated with metformin.

OBJECTIVE: To investigate the associations of metformin, serum vitamin B12, calcium supplements, and cognitive impairment in patients with diabetes. RESEARCH DESIGN AND METHODS: Participants were recruited from the Primary Research in Memory (PRIME) clinics study, the Australian Imaging, Biomarkers and Lifestyle (AIBL) study of aging, and the Barwon region of southeastern Australia. Patients with Alzheimer disease (AD) (n=480) or mild cognitive impairment (n=187) and those who were cognitively intact (n=687) were included; patients with stroke or with neurodegenerative diseases other than AD were excluded. Subgroup analyses were performed for participants who had either type 2 diabetes (n=104) or impaired glucose tolerance (n=22). RESULTS: Participants with diabetes (n=126) had worse cognitive performance than participants who did not have diabetes (n=1,228; adjusted odds ratio 1.51 [95% CI 1.03-2.21]). Among participants with diabetes, worse cognitive performance was associated with metformin use (2.23 [1.05-4.75]). After adjusting for age, sex, level of education, history of depression, serum vitamin B12, and metformin use, participants with diabetes who were taking calcium supplements had better cognitive performance (0.41 [0.19-0.92]). CONCLUSIONS: Metformin use was associated with impaired cognitive performance. Vitamin B12 and calcium supplements may alleviate metformin-induced vitamin B12 deficiency and were associated with better cognitive outcomes. Prospective trials are warranted to assess the beneficial effects of vitamin B12 and calcium use on cognition in older people with diabetes who are taking metformin.

Longitudinal analysis of serum copper and ceruloplasmin in Alzheimer's disease.

BACKGROUND: Several studies have reported that peripheral levels of copper and ceruloplasmin (CP) can differentiate patients with Alzheimer's disease (AD) from non-AD cases. The aim of this study was to determine the diagnostic value of serum copper, CP, and non-CP copper levels in a large cohort of AD subjects. METHODS: Serum copper and CP concentrations were measured at baseline and at 18-months in participants from the Australian Imaging Biomarkers and Lifestyle Study of Ageing. Cross-sectional and longitudinal analyses were conducted using both univariate and multivariate testing adjusting for age, gender, total protein, and ApoE ε4 genotype status. RESULTS: There was no significant difference in levels of serum copper or CP between the AD and healthy control groups, however, we identified a near-significant decrease in non-CP copper in the mild cognitive impairment and AD groups at baseline (p = 0.02) that was significant at 18-months (p = 0.003). CONCLUSION: Our results suggest that there may be decreased non-CP copper levels in mild cognitive impairment and AD, which is consistent with diminished copper-dependent biochemical activities described in AD.