Max Kuhn

We are moving into the age of ‘Big Data’ in biomedical research and bioinformatics. This trend could be encapsulated in this simple formula: D = S * F, where the volume of data generated (D) increases in both dimensions: the number of samples (S) and the number of sample features (F). Frequently, a typical omics classification includes redundant and irrelevant features (e.g. genes or proteins) that can result in long computation times; decrease of the model performance and the selection of… Show more

We are moving into the age of ‘Big Data’ in biomedical research and bioinformatics. This trend could be encapsulated in this simple formula: D = S * F, where the volume of data generated (D) increases in both dimensions: the number of samples (S) and the number of sample features (F). Frequently, a typical omics classification includes redundant and irrelevant features (e.g. genes or proteins) that can result in long computation times; decrease of the model performance and the selection of suboptimal features (genes and proteins) after the classification/regression step. Multiple algorithms and reviews has been published to describe all the existing methods for feature selection, their strengths and weakness. However, the selection of the correct FS algorithm and strategy constitutes an enormous challenge. Despite the number and diversity of algorithms available, the proper choice of an approach for facing a specific problem often falls in a ‘grey zone’. In this study, we select a subset of FS methods to develop an efficient workflow and an R package for bioinformatics machine learning problems. We cover relevant issues concerning FS, ranging from domain’s problems to algorithm solutions and computational tools. Finally, we use seven different proteomics and gene expression datasets to evaluate the workflow and guide the FS process. Show less

Strategic replacement of one or more hydrogen atoms with fluorine atom(s) is a common tactic to improve potency at a given target and/or to modulate parameters such as metabolic stability and pKa. Molecular weight (MW) is a key parameter in design, and incorporation of fluorine is associated with a disproportionate increase in MW considering the van der Waals radius of fluorine versus hydrogen. Herein we examine a large compound data set to understand the effect of introducing fluorine on the… Show more

Strategic replacement of one or more hydrogen atoms with fluorine atom(s) is a common tactic to improve potency at a given target and/or to modulate parameters such as metabolic stability and pKa. Molecular weight (MW) is a key parameter in design, and incorporation of fluorine is associated with a disproportionate increase in MW considering the van der Waals radius of fluorine versus hydrogen. Herein we examine a large compound data set to understand the effect of introducing fluorine on the risk of encountering P-glycoprotein mediated efflux (as measured by MDR efflux ratio), passive permeability, lipophilicity, and metabolic stability. Statistical modeling of the MDR ER data demonstrated that an increase in MW as a result of introducing fluorine atoms does not lead to higher risk of P-gp mediated efflux. Fluorine-corrected molecular weight (MWFC), where the molecular weight of fluorine has been subtracted, was found to be a more relevant descriptor. Show less

This chapter describes quantitative tools for analyzing chemical structures and relating them to assay results using statistical models. The focus is on prediction of new compounds as well as the exploratory analysis and data mining of large compound databases. Other issues related to how these analytical methods are used are discussed.

This chapter is a broad overview of the drug discovery process and areas where statistical input can have a key impact. The focus in primarily in a few key areas: target discovery, compound screening/optimization, and the characterization of important properties. Special attention is paid to working with assay data and phenotypic screens. A discussion of important skills for a nonclinical statistician supporting drug discovery concludes the chapter.

Methods: Random forest modeling of electronic medical records was used to identify variables that may facilitate earlier FM identification and diagnosis. Subjects aged ≥18 years with two or more listings of the International Classification of Diseases, Ninth Revision, (ICD-9) code for FM (ICD-9 729.1) ≥30 days apart during the 2012 calendar year were defined as cases among subjects associated with an integrated delivery network and who had one or more health care provider encounter in the… Show more

Methods: Random forest modeling of electronic medical records was used to identify variables that may facilitate earlier FM identification and diagnosis. Subjects aged ≥18 years with two or more listings of the International Classification of Diseases, Ninth Revision, (ICD-9) code for FM (ICD-9 729.1) ≥30 days apart during the 2012 calendar year were defined as cases among subjects associated with an integrated delivery network and who had one or more health care provider encounter in the Humedica database in calendar years 2011 and 2012. Controls were without the FM ICD-9 codes. Seventy-two demographic, clinical, and health care resource utilization variables were entered into a random forest model with downsampling to account for cohort imbalances (<1% subjects had FM). Importance of the top ten variables was ranked based on normalization to 100% for the variable with the largest loss in predicting performance by its omission from the model. Since random forest is a complex prediction method, a set of simple rules was derived to help understand what factors drive individual predictions.

Results: The ten variables identified by the model were: number of visits where laboratory/non-imaging diagnostic tests were ordered; number of outpatient visits excluding office visits; age; number of office visits; number of opioid prescriptions; number of medications prescribed; number of pain medications excluding opioids; number of medications administered/ordered; number of emergency room visits; and number of musculoskeletal conditions. A receiver operating characteristic curve confirmed the model's predictive accuracy using an independent test set (area under the curve, 0.810). To enhance interpretability, nine rules were developed that could be used with good predictive probability of an FM diagnosis and to identify no-FM subjects. Show less

s may guide health care providers in implementing appropriate diagnostic and management strategies.

Significant differences between the FM and no-FM cohorts were observed for nearly all the demographic, clinical, and health care resource variables, suggesting an association with FM diagnosis. These results also support use of EMR data for identifying variables associated with FM, which may help in the diagnosis and management of this condition.

Striatal-enriched protein tyrosine phosphatase (STEP) is a brain-specific tyrosine phosphatase that has been shown to de-phosphorylate several key neuronal signaling proteins, including kinases (extracellular signal-regulated kinase (ERK1/2), FYN, PYK2) and glutamate receptor subunits (N-methyl-d-aspartate receptor subtype 2B (NR2B), glutamate receptor 2 (GLUR2)). Step knock-out mice have increased phosphorylation of these substrates in the brain, with potential functional consequences in… Show more

Striatal-enriched protein tyrosine phosphatase (STEP) is a brain-specific tyrosine phosphatase that has been shown to de-phosphorylate several key neuronal signaling proteins, including kinases (extracellular signal-regulated kinase (ERK1/2), FYN, PYK2) and glutamate receptor subunits (N-methyl-d-aspartate receptor subtype 2B (NR2B), glutamate receptor 2 (GLUR2)). Step knock-out mice have increased phosphorylation of these substrates in the brain, with potential functional consequences in synaptic plasticity and cognitive tasks. It is therefore of interest to identify the molecular pathways and downstream transcriptional targets that are impacted by Step knockdown. In the present study, striatal RNA samples from Step wild-type, knock-out and heterozygous mice were hybridized to Affymetrix microarray chips and evaluated for transcriptional changes between genotypes. Pathway analysis highlighted Erk signaling and multiple pathways related to neurotrophin signaling, neuronal development and synaptic transmission. Potential genes of interest identified by microarray were confirmed by quantitative real-time polymerase chain reaction (qRT-PCR) in the cortex and hippocampus, which shared several transcriptional alterations with the striatum. In order to evaluate Step knockdown in an in vitro system, a panel of genes were evaluated using qRT-PCR in rat cortical neurons that were transduced with lentivirus expressing short hairpin RNA against Step or a non-targeting control. Our data suggest that Step has a role in the expression of immediate early genes relevant to synaptic plasticity, in both in vitro and in vivo systems. Show less

What goes on inside the black-box algorithms that turn big data into something useful? The answer, say Max Kuhn and Kjell Johnson, is statistical – so statisticians should come to the big data party.

There are a wide variety of ligand binding assay platforms available for implementation in present day bioanalytical laboratories. Selecting the platform that best suits a particular project's needs is highly dependent upon multiple assay characteristics. The active form of glucagon-like protein (GLP-1) is a biomarker of interest for type 2 diabetes (T2DM), and therefore a common target for quantitation. Previous projects requiring active GLP-1 measurements involved the use of a labor intensive… Show more

There are a wide variety of ligand binding assay platforms available for implementation in present day bioanalytical laboratories. Selecting the platform that best suits a particular project's needs is highly dependent upon multiple assay characteristics. The active form of glucagon-like protein (GLP-1) is a biomarker of interest for type 2 diabetes (T2DM), and therefore a common target for quantitation. Previous projects requiring active GLP-1 measurements involved the use of a labor intensive ELISA, spurring an investigation towards other potential assay platforms. To that end, four separate ligand binding assay formats (standard ELISA, electrochemiluminescence, Gyrolab, and Singulex) were evaluated. The platforms were compared for numerous assay parameters including dynamic range, sample volume requirements, throughput, and cost. Additionally, thirty individual donor plasmas were run with each assay as representative study samples. Although our evaluation did not show any platform that was better than others in all assay characteristics, there was one that was best in sensitivity (Singulex) and one that was best in throughput and sample volume requirements (Gyrolab). The lack of a technology that was best in all categories underscores the importance of due diligence when selecting an assay platform; there are no silver bullets, and one must take into account what is necessary for project needs and the intended use of the data. Show less

Many of the commonly observed reproductive toxicities associated with therapeutic compounds can be traced to a disruption of the steroidogenic pathway. We sought to develop an in vitro assay that would predict reproductive toxicity and be high throughput in nature. H295R cells, previously validated as having an intact and functional steroidogenic pathway, were treated with 83 known-positive and 79 known-negative proprietary and public-domain compounds. The assay measured the expression of the… Show more

Many of the commonly observed reproductive toxicities associated with therapeutic compounds can be traced to a disruption of the steroidogenic pathway. We sought to develop an in vitro assay that would predict reproductive toxicity and be high throughput in nature. H295R cells, previously validated as having an intact and functional steroidogenic pathway, were treated with 83 known-positive and 79 known-negative proprietary and public-domain compounds. The assay measured the expression of the key enzymes STAR, 3βHSD2, CYP17A1, CYP11B2, CYP19A1, CYP21A2, and CYP11A1 and the hormones DHEA, progesterone, testosterone, and cortisol. We found that a Random Forest model yielded a receiver operating characteristic area under the curve (ROC AUC) of 0.845, with sensitivity of 0.724 and specificity of 0.758 for predicting in vivo reproductive toxicity with this in vitro assay system. Show less

A set of molecules that advanced into exploratory animal toxicology studies (two species) was examined to determine what properties contributed to success in these safety studies. Compounds were rigorously evaluated across numerous safety end points and classified as “pass” if a suitable in vivo therapeutic index (TI) was achieved for advancement into regulatory toxicology studies. The most predictive end point contributing to compound survival was a predicted human efficacious concentration… Show more

A set of molecules that advanced into exploratory animal toxicology studies (two species) was examined to determine what properties contributed to success in these safety studies. Compounds were rigorously evaluated across numerous safety end points and classified as “pass” if a suitable in vivo therapeutic index (TI) was achieved for advancement into regulatory toxicology studies. The most predictive end point contributing to compound survival was a predicted human efficacious concentration (Ceff) of ≤250 nM (total drug) and ≤40 nM (free drug). This trend held across a wide range of CNS modes of action, encompassing targets such as enzymes, G-protein-coupled receptors, ion channels, and transporters. Show less

This text is intended for a broad audience as both an introduction to predictive models as well as a guide to applying them. Non-mathematical readers will appreciate the intuitive explanations of the techniques while an emphasis on problem-solving with real data across a wide variety of applications will aid practitioners who wish to extend their expertise. Readers should have knowledge of basic statistical ideas, such as correlation and linear regression analysis. While the text is biased… Show more

This text is intended for a broad audience as both an introduction to predictive models as well as a guide to applying them. Non-mathematical readers will appreciate the intuitive explanations of the techniques while an emphasis on problem-solving with real data across a wide variety of applications will aid practitioners who wish to extend their expertise. Readers should have knowledge of basic statistical ideas, such as correlation and linear regression analysis. While the text is biased against complex equations, a mathematical background is needed for advanced topics. Show less

Hyperphosphorylation of the microtubule-associated protein tau (MAPT) has been linked to several
neurodegenerative diseases including Alzheimer’s disease (AD) and frontotemporal dementia (FTD). The rTg4510 transgenic mouse model expresses a human variant of tau, P301L, which is repressible with doxycycline. Neurofibrillary tangles (NFTs), neuronal loss, and behavioral impairments can be measured in a progressive, age-dependent manner. In this study, we used in vivo volumetric MRI, localized… Show more

Hyperphosphorylation of the microtubule-associated protein tau (MAPT) has been linked to several
neurodegenerative diseases including Alzheimer’s disease (AD) and frontotemporal dementia (FTD). The rTg4510 transgenic mouse model expresses a human variant of tau, P301L, which is repressible with doxycycline. Neurofibrillary tangles (NFTs), neuronal loss, and behavioral impairments can be measured in a progressive, age-dependent manner. In this study, we used in vivo volumetric MRI, localized MRS as well as post mortem immunohistochemistry, to evaluate the effects of perinatal tau inhibition. The results demonstrated that continuous doxycycline exposure from birth until adulthood (2.5-month-old) was sufficient to delay accumulation of hyperphosphorylated tau and formation of NFTs in rTg4510 mouse brain. Volumetric MRI and histological evaluation also highlighted the absence of forebrain atrophy or neurodegeneration until at least 10 months of age. These results suggest that the most severe pathological events associated with P301L tauopathy occur at the perinatal and early postnatal stages and that providing protection against these events may significantly reduce the risks of neurodegeneration in adulthood Show less

This manuscript serves as a review of how the R language has been used in the last decade to address problems related to medicinal chemistry design. This includes the use of the R language for chemoinformatics applications and interfaces, as well as statistical modeling for ADMET and potency endpoints. Additionally, a few examples of R code are provided to demonstrate the ability of this language to make available cutting-edge statistical analysis to the medicinal chemistry design community.

Dysregulation of ceramide synthesis has been associated with metabolic disorders such as atherosclerosis and diabetes. We examined the changes in lipid homeostasis and gene expression in Huh7 hepatocytes when the synthesis of ceramide is perturbed by knocking down serine pal mitoyltransferase subunits 1, 2, and 3 (SPTLC123) or dihydroceramide desaturase 1 (DEGS1). Although knocking down all SPTLC subunits is necessary to reduce total ceramides significantly, depleting DEGS1 is sufficient to… Show more

Dysregulation of ceramide synthesis has been associated with metabolic disorders such as atherosclerosis and diabetes. We examined the changes in lipid homeostasis and gene expression in Huh7 hepatocytes when the synthesis of ceramide is perturbed by knocking down serine pal mitoyltransferase subunits 1, 2, and 3 (SPTLC123) or dihydroceramide desaturase 1 (DEGS1). Although knocking down all SPTLC subunits is necessary to reduce total ceramides significantly, depleting DEGS1 is sufficient to produce a similar outcome. Lipidomic analysis of distribution and speciation of multiple lipid classes indicates an increase in phospholipids in SPTLC123-silenced cells, whereas DEGS1 depletion leads to the accumulation of sphingolipid intermediates, free fatty acids, and diacylglycerol. When cer amide synthesis is disrupted, the transcriptional profiles indicate inhibition in biosynthetic processes, downregulation of genes involved in general endomembrane traffi cking, and upregulation of endocytosis and endosomal recycling. SPTLC123 silencing strongly affects the expression of genes involved with lipid metabolism. Changes in amino acid, sugar, and nucleotide metabolism, as well as vesicle trafficking between organelles, are more prominent in DEGS1-silenced cells. These studies are the first to provide a direct and comprehensive understanding at the lipidomic and transcriptomic levels of how Huh7 hepatocytes respond to changes in the inhibition of ceramide synthesis. Show less

Triglyceride accumulation is associated with obesity and type 2 diabetes. Genetic disruption of diacylglycerol acyltransferase 1 (DGAT1), which catalyzes the final reaction of triglyceride synthesis, confers dramatic resistance to high-fat diet induced obesity. Hence, DGAT1 is considered a potential therapeutic target for treating obesity and related metabolic disorders. However, the molecular events shaping the mechanism of action of DGAT1 pharmacological inhibition have not been fully… Show more

Triglyceride accumulation is associated with obesity and type 2 diabetes. Genetic disruption of diacylglycerol acyltransferase 1 (DGAT1), which catalyzes the final reaction of triglyceride synthesis, confers dramatic resistance to high-fat diet induced obesity. Hence, DGAT1 is considered a potential therapeutic target for treating obesity and related metabolic disorders. However, the molecular events shaping the mechanism of action of DGAT1 pharmacological inhibition have not been fully explored yet. Here, we investigate the metabolic molecular mechanisms induced in response to pharmacological inhibition of DGAT1 using a recently developed computational systems biology approach, the Causal Reasoning Engine (CRE). The CRE algorithm utilizes microarray transcriptomic data and causal statements derived from the biomedical literature to infer upstream molecular events driving these transcriptional changes. The inferred upstream events (also called hypotheses) are aggregated into biological models using a set of analytical tools that allow for evaluation and integration of the hypotheses in context of their supporting evidence. In comparison to gene ontology enrichment analysis which pointed to high-level changes in metabolic processes, the CRE results provide detailed molecular hypotheses to explain the measured transcriptional changes. CRE analysis of gene expression changes in high fat habituated rats treated with a potent and selective DGAT1 inhibitor demonstrate that the majority of transcriptomic changes support a metabolic network indicative of reversal of high fat diet effects that includes a number of molecular hypotheses such as PPARG, HNF4A and SREBPs. Finally, the CRE-generated molecular hypotheses from DGAT1 inhibitor treated rats were found to capture the major molecular characteristics of DGAT1 deficient mice, supporting a phenotype of decreased lipid and increased insulin sensitivity. Show less

Glucagon is an important hormone in the regulation of glucose homeostasis, particularly in the maintenance of euglycemia and prevention of hypoglycemia. In type 2 Diabetes Mellitus (T2DM), glucagon levels are elevated in both the fasted and postprandial states, which contributes to inappropriate hyperglycemia through excessive hepatic glucose production. Efforts to discover and evaluate glucagon receptor antagonists for the treatment of T2DM have been ongoing for approximately two decades, with… Show more

Glucagon is an important hormone in the regulation of glucose homeostasis, particularly in the maintenance of euglycemia and prevention of hypoglycemia. In type 2 Diabetes Mellitus (T2DM), glucagon levels are elevated in both the fasted and postprandial states, which contributes to inappropriate hyperglycemia through excessive hepatic glucose production. Efforts to discover and evaluate glucagon receptor antagonists for the treatment of T2DM have been ongoing for approximately two decades, with the challenge being to identify an agent with appropriate pharmaceutical properties and efficacy relative to potential side effects. We sought to determine the hepatic & systemic consequence of full glucagon receptor antagonism through the study of the glucagon receptor knock-out mouse (Gcgr-/-) compared to wild-type littermates. Show less

In this work we present a number of statistical and visualization methods derived from MIC90 data designed to aid decision-making in antibacterial drug discovery research. A statistical method known as bootstrapping was applied to MIC90 raw data to uncover data trends and a metric termed Net Percent Superior (NPS) was developed to capture a strain-by-strain analysis of analogs to enable rank-ordering of similar compounds. We also present novel methods of reporting the data using a variety of… Show more

In this work we present a number of statistical and visualization methods derived from MIC90 data designed to aid decision-making in antibacterial drug discovery research. A statistical method known as bootstrapping was applied to MIC90 raw data to uncover data trends and a metric termed Net Percent Superior (NPS) was developed to capture a strain-by-strain analysis of analogs to enable rank-ordering of similar compounds. We also present novel methods of reporting the data using a variety of visualization techniques. Furthermore, the work was cross-validated using experimental results generated with siderophore-conjugated monocarbam analogs to demonstrate the effectiveness of the various parameters and visualization techniques. The methods reported herein have been incorporated in a Scitegic Pipeline Pilot protocol to enable facile, automated generation of MIC90 analyses from experimental raw data to aid prospective medicinal chemistry design as well as retrospective analyses. Show less

Clinicopathological studies suggest that Alzheimer's disease (AD) pathology begins ~10–15 years before the resulting cognitive impairment draws medical attention. Biomarkers that can detect AD pathology in its early stages and predict dementia onset would, therefore, be invaluable for patient care and efficient clinical trial design. We utilized a targeted proteomics approach to discover novel cerebrospinal fluid (CSF) biomarkers that can augment the diagnostic and prognostic accuracy of… Show more

Clinicopathological studies suggest that Alzheimer's disease (AD) pathology begins ~10–15 years before the resulting cognitive impairment draws medical attention. Biomarkers that can detect AD pathology in its early stages and predict dementia onset would, therefore, be invaluable for patient care and efficient clinical trial design. We utilized a targeted proteomics approach to discover novel cerebrospinal fluid (CSF) biomarkers that can augment the diagnostic and prognostic accuracy of current leading CSF biomarkers (Aβ42, tau, p-tau181). Show less

Inhibition of phosphodiesterase 10A (PDE10A) promotes cyclic nucleotide signaling, increases striatal activation and decreases behavioral activity. Enhanced cyclic nucleotide signaling is a well-establish route to producing changes in gene expression. We hypothesized that chronic suppression of PDE10A activity would have significant effects on gene expression in the striatum. A comparison of the expression profile of PDE10A knockout mice (KO) and wild-type (WT) mice following chronic PDE10A… Show more

Inhibition of phosphodiesterase 10A (PDE10A) promotes cyclic nucleotide signaling, increases striatal activation and decreases behavioral activity. Enhanced cyclic nucleotide signaling is a well-establish route to producing changes in gene expression. We hypothesized that chronic suppression of PDE10A activity would have significant effects on gene expression in the striatum. A comparison of the expression profile of PDE10A knockout mice (KO) and wild-type (WT) mice following chronic PDE10A inhibition revealed altered expression of 19 overlapping genes with few significant changes outside the striatum or following administration of a PDE10A inhibitor to KO animals. Chronic inhibition of PDE10A produced up-regulation of mRNAs encoding genes that included prodynorphin, synaptotagmin10, phosphodiesterase 1C (PDE1C), glutamate decarboxylase 1 (GAD67), diacylglycerol O-acyltransferase (DGAT2) and a down regulation of mRNA encoding choline acetyltransferase (ChAT) and Kv1.6, suggesting long-term suppression of the PDE10A enzyme is consistent with altered striatal excitability and potential utility as a antipsychotic therapy. Additionally, upregulation of mRNA encoding histone H3 and downregulation of histone deacetylase 4, follistatin and claspin mRNAs suggests activation of molecular cascades capable of neuroprotection. We utilized lentiviral delivery of CRE-luciferase reporter constructs into the striatum and live animal imaging of TP-10 induced luciferase activity to further demonstrate PDE10 inhibition results in CRE-mediated transcription. Consistent with potential neuroprotective cascades, we also demonstrate phosphorylation of mitogen- and stress-activated kinases 1 (MSK1) and histone H3 in vivo following TP-10 treatment. The observed changes in signaling and gene expression are predicted to provide neuroprotective effects in models of Huntington's Disease. Show less

Interventions for T2DM have in part aimed to mimic exercise. Here, we have compared the independent and combined effects of a PPARδ agonist and endurance training mimetic (GW501516) and a myostatin antibody and resistance training mimetic (PF-879) on metabolic and performance outcomes in obese insulin resistant mice.

Altered levels of cerebrospinal fluid (CSF) peptides related to Alzheimer’s disease (AD) are associated with pathologic AD diagnosis, although cognitively normal subjects can also have abnormal levels of these AD biomarkers. To identify novel CSF biomarkers that distinguish pathologically confirmed AD from cognitively normal subjects and patients with other neurodegenerative disorders, we collected antemortem CSF samples from 66 AD patients and 25 patients with other neurodegenerative dementias… Show more

Altered levels of cerebrospinal fluid (CSF) peptides related to Alzheimer’s disease (AD) are associated with pathologic AD diagnosis, although cognitively normal subjects can also have abnormal levels of these AD biomarkers. To identify novel CSF biomarkers that distinguish pathologically confirmed AD from cognitively normal subjects and patients with other neurodegenerative disorders, we collected antemortem CSF samples from 66 AD patients and 25 patients with other neurodegenerative dementias followed longitudinally to neuropathologic confirmation, plus CSF from 33 cognitively normal subjects. We measured levels of 151 novel analytes via a targeted multiplex panel enriched in cytokines, chemokines and growth factors, as well as established AD CSF biomarkers (levels of Aβ42, tau and p-tau181). Two categories of biomarkers were identified: (1) analytes that specifically distinguished AD (especially CSF Aβ42 levels) from cognitively normal subjects and other disorders; and (2) analytes altered in multiple diseases (NrCAM, PDGF, C3, IL-1α), but not in cognitively normal subjects. A multi-prong analytical approach showed AD patients were best distinguished from non-AD cases (including cognitively normal subjects and patients with other neurodegenerative disorders) by a combination of traditional AD biomarkers and novel multiplex biomarkers. Six novel biomarkers (C3, CgA, IL-1α, I-309, NrCAM and VEGF) were correlated with the severity of cognitive impairment at CSF collection, and altered levels of IL-1α and TECK associated with subsequent cognitive decline in 38 longitudinally followed subjects with mild cognitive impairment. In summary, our targeted proteomic screen revealed novel CSF biomarkers that can improve the distinction between AD and non-AD cases by established biomarkers alone. Show less