Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Parallel Algorithms”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,369 records · Page 76Linked to original sources
PubMed1993

Spontaneous burst firing in cat primary auditory cortex: age and depth dependence and its effect on neural interaction measures.

1. Neural activity was recorded with two independent electrodes separated by 0.5-2 mm, aligned in parallel, and advanced perpendicular to the surface of the cat auditory cortex. Because the experiments were part of a study into laminar interaction the difference in recording depths for the two independently movable electrodes was never > 100 microns. Multi-unit activity on each electrode was separated on-line into single-unit spike-trains with a maximum variance spike sorting algorithm. Off-line controls on the quality of the spike-train separation were routinely performed. The first aim of this study was to describe the age dependence of spontaneous burst firing and to explore if and how it could be explained by age dependent changes in firing rate. The second aim was to investigate a potential layer dependence on burst firing. The third aim was to describe the effect of burst-removal procedures on the shape, strength, and width of the cross-correlogram and to investigate whether an age dependence in burst firing might account for the previously reported age dependence in correlation strengths. 2. Recordings were made from 237 single units from primary auditory cortex in nine adult cats and from 67 units in seven kittens age 10-52 days. The incidence of burst firing as a function of firing rate, age and depth of recording and unit characteristic frequency was investigated. In addition the effect of burst firing on the strength and width of the central peak in 471 neural pair correlograms was analyzed. 3. Burst firing could be distinguished at many different time scales; bursts lasting of the order of 10 s contained bursts with durations of the order of 1 s, which in turn contained bursts of 30-50-ms duration. The analysis in this paper was restricted to the short-duration bursts. 4. Burst firing on the short-time scale of 50 ms was characterized by relatively well defined intervals between the first two spikes (3-15 ms) followed by higher-order intervals with large spread (range 4-50 ms) but with increasing modal interval value. The typical adult five-spike burst template featured spikes at 0, 3.3, 14.6, 27.2, and 34.8 ms. Burst with fewer spikes showed larger intervals between the first three spikes. 5. The probability of occurrence of isolated spikes, pairs, triplets, etc. showed a power-law dependence on firing rate with a coefficient that was significantly lower than expected under Poisson firing conditions.(ABSTRACT TRUNCATED AT 400 WORDS)

Action Potentials↗
PubMed2006

Single-cell analysis of glucocorticoid receptor action reveals that stochastic post-chromatin association mechanisms regulate ligand-specific transcription.

The glucocorticoid receptor (GR) dynamically interacts with response elements in the mouse mammary tumor virus (MMTV) promoter to regulate steroid-dependent transcription. In a clonal mammary carcinoma cell line containing a tandem array of MMTV promoter-reporter gene cassettes integrated at a single genomic locus, direct binding of a green fluorescent protein (GFP)-GR fusion protein to the MMTV regulatory elements can be observed in living cells. After ligand treatment, MMTV-dependent transcription in individual cells was detected by RNA fluorescence in situ hybridization (FISH). High-resolution fluorescence images were acquired from large numbers of randomly selected cells. Images were analyzed with a novel automated computer algorithm, measuring the RNA FISH signal and the relative GFP-GR fluorescence intensity at the MMTV array for each cell. Although dexamethasone increased the mean RNA FISH signal approximately 10-fold, RU486 produced only about a 2-fold induction, as expected for this mixed antagonist. For all treatment conditions, the relative GFP-GR fluorescence at the array for the averaged cells paralleled the RNA FISH measurements, suggesting that image analysis accurately detected an increase in steady-state GR association with the MMTV array that was responsible for the increase in transcriptional activity. The antagonist-dependent decreases in GR association with the MMTV promoter were confirmed by chromatin immunoprecipitation experiments, supporting the image analysis results. A pronounced cell-to-cell variability was observed in RNA FISH signal and GR-MMTV association within treatment groups. We observed a nonlinear relationship between GR-MMTV association and RNA FISH in individual cells, indicating that differences in GR-MMTV interaction account for some, but not all, of the transcriptional heterogeneity between individual cells. In selected cell subpopulations with equal levels of GR-MMTV association, there was a decrease in RNA FISH signal with RU486 treatment compared with dexamethasone treatment. These results indicate that stochastic events occurring after GR-promoter association, such as the actions of chromatin remodeling complexes or other cofactors, change in a ligand-dependent manner and regulate heterogeneous transcription in individual cells.

Animals↗
PubMed2006

Statistical representation of mean diffusivity and fractional anisotropy brain maps of normal subjects.

PURPOSE: To create diffusion tensor atlases from echo planar imaging (EPI) images acquired at 3 T in 10 normal subjects. MATERIALS AND METHODS: Data from 10 right-handed healthy adult volunteers (mean age of 31 +/- 3 years; eight males) were acquired using a 3.0-T scanner. Geometric distortion artifacts correction was accomplished by combining parallel acquisition to reduce the distortion as well as postprocessing by registration to a geometrically accurate T2-weighted fast-spin-echo image. This reduced distortions to within a voxel for most of the internal structures of the brain. The apparent diffusion coefficient (ADC) and fractional anisotropy (FA) atlases were created by warping images using an iterative optical-flow-based local deformation algorithm that used two channels of data: ADC and FA. RESULTS: A three-dimensional distance measure was used to evaluate the accuracy of the registration algorithm with contours defined on two structures: the corpus callosum and cerebellum. The average three-dimensional distance value for the nine subjects (with the 10th as the reference) was 0.2 mm for the corpus callosum and 1.2 mm for the cerebellum. CONCLUSION: A high-resolution, diffusion MR atlas with full brain coverage was developed. Additionally, maps of the SD of the diffusion indices were also generated to provide an estimate of the variance within a normal population. Active shape and texture models were also generated for the corpus callosum as an alternate method of representing the variance in morphology and diffusion indices.

Adult↗
PubMedDate not supplied

Knowledge-assisted recognition of cluster boundaries in gene expression data.

BACKGROUND AND MOTIVATION: DNA microarray technology has made it possible to determine the expression levels of thousands of genes in parallel under multiple experimental conditions. Genome-wide analyses using DNA microarrays make a great contribution to the exploration of the dynamic state of genetic networks, and further lead to the development of new disease diagnosis technologies. An important step in the analysis of gene expression data is to classify genes with similar expression patterns into the same groups. To this end, hierarchical clustering algorithms have been widely used. Major advantages of hierarchical clustering algorithms are that investigators do not need to specify the number of clusters in advance and results are presented visually in the form of a dendrogram. However, since traditional hierarchical clustering methods simply provide results on the statistical characteristics of expression data, biological interpretations of the resulting clusters are not easy, and it requires laborious tasks to unveil hidden biological processes regulated by members in the clusters. Therefore, it has been a very difficult routine for experts. OBJECTIVE: Here, we propose a novel algorithm in which cluster boundaries are determined by referring to functional annotations stored in genome databases. MATERIALS AND METHODS: The algorithm first performs hierarchical clustering of gene expression profiles. Then, the cluster boundaries are determined by the Variance Inflation Factor among the Gene Function Vectors, which represents distributions of gene functions in each cluster. Our algorithm automatically specifies a cutoff that leads to functionally independent agglomerations of genes on the dendrogram derived from similarities among gene expression patterns. Finally, each cluster is annotated according to dominant gene functions within the respective cluster. RESULTS AND CONCLUSIONS: In this paper, we apply our algorithm to two gene expression datasets related to cell cycle and cold stress response in budding yeast Saccharomyces cerevisiae. As a result, we show that the algorithm enables us to recognize cluster boundaries characterizing fundamental biological processes such as the Early G1, Late G1, S, G2 and M phases in cell cycles, and also provides novel annotation information that has not been obtained by traditional hierarchical clustering methods. In addition, using formal cluster validity indices, high validity of our algorithm is verified by the comparison through other popular clustering algorithms, K-means, self-organizing map and AutoClass.

Algorithms↗
PubMed2005

Integration of a handheld based anaesthesia rounding system into an anaesthesia information management system.

BACKGROUND: At the University Hospital Giessen, an anesthesia information management system (AIMS) is used for online record keeping of perioperative patient care, but preoperative anaesthesia assessments were still being recorded on paper and subsequently entered into the AIMS. Personal digital assistants (PDAs) seem to be useful instruments to establish a seamless digital anesthesiological documentation. OBJECTIVES: We decided to implement a solution for direct integration of data gathered during the preoperative assessment into the existing data management infrastructure. Parallel to the development of the system, we surveyed the future users to match their wishes and needs as far as possible. SYSTEM DESCRIPTION: A C program embedding the preoperative AIMS' data fields was developed. Data alignment with the Hospital information system (HIS) is controlled by a Java desktop software. The anaesthesiologist completes the available fields at the patient's bedside following the same algorithm and integrity check as the PC version. STATUS REPORT: Overall, 68% of the surveyed physicians supported the implementation of the system. The PDA solution has been available since May 2002. Data replication into the handheld and integration of mobile collected data into the AIMS generally work without problems. The HIS interconnection software converts the PDA file into the AIMS format for further processing. DISCUSSION: The preoperative anaesthetic assessment is a standardised task well suitable for conversion to an electronic data storage medium. Changing from redundant data entry in the OR to direct electronic recording at the patient's bedside seems simply logical. Handheld computers are inexpensive, flexible gadgets to realize this.

Anesthesia↗
PubMed1998

A new quantitative criterion to distinguish between alpha/beta and alpha+beta proteins (domains).

According to the statistical analysis, it is shown that the differences of the content of alpha-helix and beta-strand between alpha/beta and alpha+beta proteins are of statistical significance. Based on the secondary structure content and the percentage of parallel or anti-parallel strands, any mixed alphabeta protein can be represented by a point in a three-dimensional prism. The distribution of the mapping points for 79 mixed alphabeta proteins (domains), of which 26 are class alpha/beta and 53 are class alpha+beta, shows that the two kinds of points are situated at distinct regions roughly. A new quantitative criterion based on the Fisher discriminant algorithm is proposed to distinguish between the alpha/beta and alpha+beta proteins (domains). Of the 79 proteins 77 are correctly classified (97.5%). As a stringent cross-validation test, the jackknife test shows that of the 79 proteins 77 are correctly classified. The jackknife test accuracy is still 97.5%. These figures indicate the self-consistence and the extrapolating effectiveness of the new quantitative criterion. Applying the new criterion to reclassify the alpha/beta and alpha+beta proteins (domains) in SCOP is also discussed. It is hoped that the new quantitative criterion will be useful for the development of protein classification databases.

Algorithms↗
PubMed2001

A high-resolution phantom for MRI.

Assessment of spatial resolution is an important step to test the performance of new sequence techniques-especially ultrafast techniques with dedicated k-space trajectories or interpolation algorithms. Measurement of the modulation transfer function (MTF) is a rather difficult procedure, but using suitable resolution phantoms allows a simple visual evaluation of spatial resolution. In contrast to commonly used test objects with a very restricted number of resolution patterns we developed a phantom containing resolution patterns from 0.1 to 1.5 mm in steps of 0.1 mm. One resolution pattern consists of five parallel Plexiglas strips with the distance of the strips being equal to their thickness. Together with a Plexiglas cuboid the resolution patterns are mounted on a Plexiglas plate on the bottom of the cylindrical phantom. An aqueous solution of manganese chloride is used to fill the phantom. High resolution cross sections (pixel size: 50 microm) through the resolution patterns were measured to confirm the correct dimensions of the phantom. To verify the appropriateness of the 0.1 and 0.2 mm stacks micro-CT images with a pixel size of 25 microm were acquired additionally for both patterns. Besides visual inspection evaluation of the profile function of signal intensity across the stacks demonstrates that the resolution patterns are sufficiently correct. T(1)-weighted SE sequences with slightly different pixel sizes as well as T(1)- and T(2*)- weighted gradient echo sequences were applied to demonstrate some possible applications of this phantom. In conclusion, the proposed phantom is well suited to assess the spatial resolution qualitatively (i.e., visually) and quantitatively over a wide range in steps of 0.1 mm.

Humans↗
PubMed2006

Protein fold recognition using segmentation conditional random fields (SCRFs).

Protein fold recognition is an important step towards understanding protein three-dimensional structures and their functions. A conditional graphical model, i.e., segmentation conditional random fields (SCRFs), is proposed as an effective solution to this problem. In contrast to traditional graphical models, such as the hidden Markov model (HMM), SCRFs follow a discriminative approach. Therefore, it is flexible to include any features in the model, such as overlapping or long-range interaction features over the whole sequence. The model also employs a convex optimization function, which results in globally optimal solutions to the model parameters. On the other hand, the segmentation setting in SCRFs makes their graphical structures intuitively similar to the protein 3-D structures and more importantly provides a framework to model the long-range interactions between secondary structures directly. Our model is applied to predict the parallel beta-helix fold, an important fold in bacterial pathogenesis and carbohydrate binding/cleavage. The cross-family validation shows that SCRFs not only can score all known beta-helices higher than non-beta-helices in the Protein Data Bank (PDB), but also accurately locates rungs in known beta-helix proteins. Our method outperforms BetaWrap, a state-of-the-art algorithm for predicting beta-helix folds, and HMMER, a general motif detection algorithm based on HMM, and has the additional advantage of general application to other protein folds. Applying our prediction model to the Uniprot Database, we identify previously unknown potential beta-helices.

Algorithms↗
PubMedDate not supplied

The myocardial microangiopathy in human and experimental diabetes mellitus. (A microscopic, ultrastructural, morphometric and computer-assisted symbolic-logic analysis).

The following microscopical aspects were found in the small intramural arteries in the myocardium of 30 diabetic patients: endothelial proliferations with focal protuberances leading to partial narrowing of the lumen, increased thickness of the arterial wall due to fibrosis and accumulations of neutral mucopolysaccharides: alteration of elastic fibres. Morphometrically, the arterial wall thickness and the arterial diameter were increased whereas the arterial density decreased in the diabetic heart. In 25 rats with streptozotocin-induced diabetes the small intramyocardial arteries were investigated at 11 to 40 weeks of diabetic state. Using morphometrical analysis a constant increase of arterial wall thickness paralleling the diabetes duration was found. Microscopically, the lesions consist in endothelial proliferation with bridging across the vascular lumen and slight perivascular and diffuse fibrosis. Ultrastructurally, the capillary basal lamina was thickened in the diabetic myocardium. In order to investigate the morphometrical data we used symbolic-logic as a decision method, by applying an original computer program based on the Quine-McCluskey algorithm. All our results together with the final symbolic-logic expression suggest that damage of the small intramyocardial arteries plays an important role in the pathogenesis of diabetic cardiomyopathy.

Aged↗
PubMed1991

A platform for biological sequence comparison on parallel computers.

We have written two programs for searching biological sequence databases that run on Intel hypercube computers. PSCANLIB compares a single sequence against a sequence library, and PCOMPLIB compares all the entries in one sequence library against a second library. The programs provide a general framework for similarity searching; they include functions for reading in query sequences, search parameters and library entries, and reporting the results of a search. We have isolated the code for the specific function that calculates the similarity score between the query and library sequence; alternative searching algorithms can be implemented by editing two files. We have implemented the rapid FASTA sequence comparison algorithm and the more rigorous Smith-Waterman algorithm within this framework. The PSCANLIB program on a 16 node iPSC/2 80386-based hypercube can compare a 229 amino acid protein sequence with a 3.4 million residue sequence library in approximately 16 s with the FASTA algorithm. Using the Smith-Waterman algorithm, the same search takes 35 min. The PCOMPLIB program can compare a 0.8 million amino acid protein sequence library with itself in 5.3 min with FASTA on a third-generation 32 node Intel iPSC/860 hypercube.

Algorithms↗
PubMed2005

Continuous arterial spin labeling perfusion measurements using single shot 3D GRASE at 3 T.

Single shot 3D GRASE is less sensitive to field inhomogeneity and susceptibility effects than gradient echo based fast imaging sequences while preserving the acquisition speed. In this study, a continuous arterial spin labeling (CASL) pulse was added prior to the single shot 3D GRASE readout and quantitative perfusion measurements were carried out at 3 T, at rest and during functional activation. The sequence performance was evaluated by comparison with a CASL sequence with EPI readout. It is shown that perfusion measurements using CASL GRASE can be performed safely on humans at 3 T without exceeding the current RF power deposition limits. The maps of resting cerebral blood flow generated from the GRASE images are comparable to those obtained with the 2D EPI readout, albeit with better coverage in the orbitofrontal cortex. The sequence proved effective for functional imaging, yielding time series of images with improved temporal SNR with respect to EPI and group activation maps with increased significance levels. The method was further improved using parallel imaging techniques to provide increased spatial resolution and better separation of the gray-white matter cerebral blood flow maps.

Adult↗
PubMed1997

Electron tomography of large, multicomponent biological structures.

Electron tomography is an extremely useful method for deriving three-dimensional structure from electron microscope images. The application of this technique to the reconstruction of large, complex structures such as mitochondria is described in conjunction with several tools for segmentation, measurement, classification, and visualization. In addition, the use of massively parallel computers to perform the tomographic reconstruction efficiently using R-weighted backprojection or iterative techniques is described.

Algorithms↗
PubMed2003

Quantitative characterization of visual response properties in the mouse dorsal lateral geniculate nucleus.

We present a quantitative analysis of the visual response properties of single neurons in the dorsal lateral geniculate nucleus (dLGN) of wild-type C57Bl/6J mice. Extracellular recordings were made from single dLGN cells in mice under halothane and nitrous oxide anesthesia. After mapping the receptive fields (RFs) of these cells using reverse correlation of responses to flashed square stimuli, we used sinusoidal gratings to describe their linearity of spatial summation, spatial frequency tuning, temporal frequency tuning, and contrast response characteristics. All cells in our sample had RFs dominated by a single, roughly circular "center" mechanism that responded to either increases (ON-center) or decreases (OFF-center) in stimulus luminance, and almost all cells passed a modified null test for linearity of spatial summation. A difference of Gaussians model was used to relate spatial frequency tuning to the spatial properties of cells' RFs, revealing that mouse dLGN cells have large RFs (center diameter approximately 11 degrees) and correspondingly poor spatial resolution (approximately 0.2c/degree). Temporally, most cells in the mouse dLGN respond best to stimuli of approximately 4 Hz. We looked for evidence of parallel processing in the mouse dLGN and found it only in a functional difference between ON- and OFF-center cells: ON-center cells were more sensitive to stimulus contrast than their OFF-center neighbors.

Algorithms↗
PubMed1988

Computer-aided sleep staging in clinical environment.

In the present paper we will comment on how we have been implementing computer-aided sleep staging (CAS) in the Department of Psychiatry, Erasmus Hospital (Free University of Brussels). Major features of our CAS include real-time stage estimation, full report and hypnogram available in the morning, and evolution of state variables throughout the night (e.g. electroencephalographic spectral characteristics, eye movement density, myographic activity). Parallel, we report an extensive validation study of this system on 50 subjects divided in two groups of 25 patients with psychiatric diseases, and 25 normal subjects.

Adult↗
PubMed2004

Modeling 3-D compliant blood flow with FOSLS.

Blood flow in large vessels is typically modeled using the Navier-Stokes equations for the fluid domain and elasticity equations for the vessel wall. As the wall deforms, additional complications are introduced because the shape of the fluid domain changes, necessitating the use of a re-mapping or re-griding process for the fluid region. Typically, this system (fluid, solid, mapping) is solved using an iterative approach in which the fluid, elastic, and mapping equations are solved in series until the iterations converge. We present a new approach based on multilevel minimization of the finite element approximation error using a least-squares (LS) norm. This approach allows for minimization of the error for the entire system or in selected parts. The multilevel LS approach overcomes many shortcomings of standard techniques. Most notably, the computational cost of solving the problem increases linearly with the degrees of freedom and the associated least-squares functional provides an a posteriori error measure. This paper compares the LS finite element approach to other popular numerical methods, specifically, the commercial package CFD-ACE. The focus of the comparison is on accuracy, computational cost, scalability (both parallel and serial), and flexibility. We show that the multilevel LS finite element approach scales optimally (i.e., linearly in serial environments), while the other methods degrade substantially as the problem size increases.

Algorithms↗
PubMed2003

The shortest common supersequence problem in a microarray production setting.

MOTIVATION: During microarray production, several thousands of oligonucleotides (short DNA sequences) are synthesized in parallel, one nucleotide at a time. We are interested in finding the shortest possible nucleotide deposition sequence to synthesize all oligos in order to reduce production time and increase oligo quality. Thus we study the shortest common super-sequence problem of several thousand short strings over a four-letter alphabet. RESULTS: We present a statistical analysis of the basic ALPHABET-LEFTMOST approximation algorithm, and propose several practical heuristics to reduce the length of the super-sequence. Our results show that it is hard to beat ALPHABET-LEFTMOST in the microarray production setting by more than 2 characters, but these savings can improve overall oligo quality by more than four percent. AVAILABILITY: Source code in C may be obtained by contacting the author, or from http://oligos.molgen.mpg.de.

Algorithms↗
PubMed2002

High-precision measurement of the vocal fold length and vibratory amplitudes.

OBJECTIVE/HYPOTHESIS: Standard laryngoscopy suffers from a lack of information about the actual size of the observed objects (i.e., vocal fold length and oscillating amplitudes). However, there is much interest in absolute measures for both clinical and research purposes. Therefore, a laser projection device has been developed that enables the precise determination of absolute units in endoscopic investigation during respiration and phonation. STUDY DESIGN: An experimental study in which 9 adults underwent high-speed endoscopy combined with a laser projection device. METHODS: The projection system consists of two parallel laser beams with a distance of 3.8 mm. It is mounted on the tip of a rigid endoscope, which is attached to a digital high-speed camera During development and design, examination situations were taken into account. Two laser spots are projected onto the vocal folds and enable the definition of a metric scale within the endoscopic image. Knowledge-based image processing algorithms were used for evaluation. RESULTS: First measurements of the vocal fold length during phonation were performed in a group of nine healthy male students. The determination of glottal length during phonation agrees with former results. Quantifying vocal fold velocities in absolute units makes it possible to estimate the initial collision forces. CONCLUSIONS: The presented laser projection system allows the determination of absolute measures in the larynx. Because of the simple functional principle, the system is open for use without digital high-speed recording as well. Absolute units may also be helpful in voice diagnosis and in monitoring during voice therapy.

Adult↗
PubMed2005

Initiating insulin therapy in type 2 Diabetes: a comparison of biphasic and basal insulin analogs.

OBJECTIVE: Safety and efficacy of biphasic insulin aspart 70/30 (BIAsp 70/30, prebreakfast and presupper) were compared with once-daily insulin glargine in type 2 diabetic subjects inadequately controlled on oral antidiabetic drugs (OADs). RESEARCH DESIGN AND METHODS: This 28-week parallel-group study randomized 233 insulin-naive patients with HbA(1c) values >/=8.0% on >1,000 mg/day metformin alone or in combination with other OADs. Metformin was adjusted up to 2,550 mg/day before insulin therapy was initiated with 5-6 units BIAsp 70/30 twice daily or 10-12 units glargine at bedtime and titrated to target blood glucose (80-110 mg/dl) by algorithm-directed titration. RESULTS: A total of 209 subjects completed the study. At study end, the mean HbA(1c) value was lower in the BIAsp 70/30 group than in the glargine group (6.91 +/- 1.17 vs. 7.41 +/- 1.24%, P < 0.01). The HbA(1c) reduction was greater in the BIAsp 70/30 group than in the glargine group (-2.79 +/- 0.11 vs. -2.36 +/- 0.11%, respectively; P < 0.01), especially for subjects with baseline HbA(1c) >8.5% (-3.13 +/- 1.63 vs. -2.60 +/- 1.50%, respectively; P < 0.05). More BIAsp 70/30-treated subjects reached target HbA(1c) values than glargine-treated subjects (HbA(1c) </=6.5%: 42 vs. 28%, P < 0.05; HbA(1c) <7.0%: 66 vs. 40%, P < 0.001). Minor hypoglycemia (episodes/year) was greater in the BIAsp 70/30 group than in the glargine group (3.4 +/- 6.6 and 0.7 +/- 2.0, respectively; P < 0.05). Weight gain and daily insulin dose at study end were greater for BIAsp 70/30-treated subjects than for glargine-treated subjects (weight gain: 5.4 +/- 4.8 vs. 3.5 +/- 4.5 kg, P < 0.01; insulin dose: 78.5 +/- 39.5 and 51.3 +/- 26.7 units/day, respectively). CONCLUSIONS: In subjects with type 2 diabetes poorly controlled on OADs, initiating insulin therapy with twice-daily BIAsp 70/30 was more effective in achieving HbA(1c) targets than once-daily glargine, especially in subjects with HbA(1c) >8.5%.

Adult↗