SEARCH · PubMed Health

Results for “sparsity”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

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

At least 19 recordsLinked to original sources

PubMed2025

Sparse deconvolution of cell type medleys in spatial transcriptomics.

Mapping cell distributions across spatial locations with whole-genome coverage is essential for understanding cellular responses and signaling However, current deconvolution models aim to estimate the proportions of distinct cell types in each spatial transcriptomics spot by integrating reference single-cell data. These models often assume strong overlap between the reference and spatial datasets, neglecting biology-grounded constraints such as sparsity and cell-type variations, as well as technical sparsity. As a result, these methods rely on over-permissive algorithms that ignore given constraints leading to inaccurate predictions, particularly in heterogeneous or unmatched datasets. We introduce Weight-Induced Sparse Regression (WISpR), a machine learning algorithm that integrates spot-specific hyperparameters and sparsity-driven modeling. Unlike conventional approaches that neglect biology-grounded constraints, WISpR accurately predicts cell-type distributions while preserving biological coherence, i.e., spatially and functionally consistent cell-type localization, even in unmatched datasets. Benchmarking against five alternative methods across ten datasets, WISpR consistently outperformed competitors and predicted cellular landscapes in both normal and cancerous tissues. By leveraging sparse cell-type arrangements, WISpR provides biologically informed, high-resolution cellular maps. Its ability to decode tissue organization in both healthy and diseased states highlights WISpR's practical utility for spatial transcriptomics, particularly in challenging settings involving noise, sparsity, or reference mismatches.

Humans↗

PubMed2026

ZILA-SRM: a probabilistic framework with zero-inflated latent models for robust strain reconstruction from metagenomes.

UNLABELLED: Resolving bacterial strain diversity from shotgun metagenomic data is fundamental to understanding intra-host evolution, transmission dynamics, and phenotypic heterogeneity. However, current probabilistic approaches face a severe "identifiability limit" when disentangling highly similar genomes. Under high-noise conditions, sequencing errors, coverage overdispersion, and collinearity confound standard expectation-maximization algorithms, resulting in overfitting and spurious "ghost" strains. Here, we introduce zero-inflated latent allocation for strain reconstruction from metagenomes with adaptive sparsity regularization (ZILA-SRM) to overcome this barrier through three innovations. First, we integrate a zero-inflated Poisson mixture model to decouple "structural zeros" (true strain absence) from "sampling zeros" (stochastic dropout), addressing overdispersion in standard Poisson-based tools. Second, we impose a convex adaptive sparsity regularization penalty that leverages biological sparsity priors to shrink noise artifacts dynamically. Third, we implement a graph-theoretic refinement step using maximal clique enumeration to resolve haplotype collinearity. Benchmarking against StrainFinder and MixtureS on 702 synthetic data sets shows that ZILA-SRM achieves a 20% improvement in precision in high-complexity scenarios while maintaining over 80% recall for minor variants at 0.5% abundance. Re-analysis of deep-sequencing data from 195 Mycobacterium tuberculosis clinical samples reveals cryptic low-abundance drug-resistant variants in 12% of patients, including a minor clone carrying the rpoB S450L mutation. Furthermore, application to skin microbiome data sets further reveals a strong negative correlation between dominant Staphylococcus aureus and Staphylococcus epidermidis strains, providing genomic evidence for competitive exclusion. These findings establish ZILA-SRM as a robust tool for resolving strain-level diversity in complex metagenomes. IMPORTANCE: Understanding microbial communities at the strain level is critical because closely related strains can differ dramatically in traits such as drug resistance, virulence, and ecological interactions. However, resolving individual strains from metagenomic sequencing data remains difficult, especially when strains are highly similar or present at low abundance. As a result, biologically meaningful diversity is often obscured or misinterpreted as noise. In this study, we introduce a new framework that improves the reliability of strain reconstruction from complex metagenomic data. By reducing false-positive strain detection while preserving sensitivity to rare variants, our approach enables more accurate characterization of microbial populations. This improved resolution reveals previously hidden subpopulations in clinical and microbiome datasets, providing clearer insights into microbial evolution, competition, and the emergence of clinically relevant traits such as antibiotic resistance.

Metagenomics↗

PubMed2025

A novel high-dimensional model for identifying regional DNA methylation QTLs.

Varying coefficient models offer the flexibility to learn the dynamic changes of regression coefficients. Despite their good interpretability and diverse applications, in high-dimensional settings, existing estimation methods for such models have important limitations. For example, we routinely encounter the need for variable selection when faced with a large collection of covariates with nonlinear/varying effects on outcomes, and no ideal solutions exist. One illustration of this situation could be identifying a subset of genetic variants with local influence on methylation levels in a regulatory region. To address this problem, we propose a composite sparse penalty that encourages both sparsity and smoothness for the varying coefficients. We present an efficient proximal gradient descent algorithm that scales to high-dimensional predictor spaces, providing sparse solutions for the varying coefficients. A comprehensive simulation study has been conducted to evaluate the performance of our approach in terms of estimation, prediction and selection accuracy. We show that the inclusion of smoothness control yields much better results over sparsity-only approaches. An adaptive version of the penalty offers additional performance gains. We further demonstrate the utility of our method in identifying regional mQTLs from asymptomatic samples in the CARTaGENE cohort. The methodology is implemented in the R package sparseSOMNiBUS, available on GitHub.

Humans↗

PubMed2025

MyESL: A Software for Evolutionary Sparse Learning in Molecular Phylogenetics and Genomics.

Evolutionary sparse learning uses supervised machine learning to build evolutionary models where genomic sites loci are parameters. It uses the Least Absolute Shrinkage and Selection Operator with bi-level sparsity to connect a specific phylogenetic hypothesis with sequence variation across genomic loci. The MyESL software addresses the need for open-source tools to perform evolutionary sparse learning analyses, offering features to preprocess input phylogenomic alignments, post-process output models to generate molecular evolutionary metrics, and make Least Absolute Shrinkage and Selection Operator regression adaptable and efficient for phylogenetic trees and alignments. The core of MyESL, which constructs models with logistic regressions using bi-level sparsity, is written in C++. Its input data preprocessing and result post-processing tools are developed in Python. Compared to other tools, MyESL is more computationally efficient and provides evolution-friendly inputs and outputs. These features have already enabled the use of MyESL in two phylogenomic applications, one to identify outlier sequences and fragile clades in inferred phylogenies and another to build genetic models of convergent traits. In addition to the use in a Python environment, MyESL is available as a standalone executable compatible across multiple platforms, which can be directly integrated into scripts and third-party software. The source code, executable, and documentation for MyESL are openly accessible at https://github.com/kumarlabgit/MyESL.

Phylogeny↗

PubMed2026

Identifying fundamental gaps in functional metagenomics: a step towards unlocking microbiome research potential.

Incomplete functional annotation limits biological interpretation in microbiome studies and their translational potential. Poor annotation arises from multiple causes, with incomplete gene-protein-reaction mapping being one tractable yet under-examined contributor. We address this gap by developing a comprehensive hierarchical framework that systematically integrates gene families in UniRef, proteins in UniProt, and metabolic reactions in MetaCyc and BioCyc through UniProtKB accession, EC number, and Pfam-domain matching. Applied to a human gut metagenome dataset via HUMAnN3, our MetaCyc-based mapping recovers up to 2.3-fold more unique reaction identifiers than the default pipeline and increases reaction prevalence across samples from ≈32% to 52% core reactions, addressing the data sparsity that limits statistical and machine-learning applications in microbiome research. Biological plausibility for the tested functions was supported by positive and negative controls: gut-microbial hormone-metabolism reactions previously linked to this dataset were recovered, while vertebrate-specific hormone-metabolism reactions remained correctly undetected. These gains derive from systematic database integration alone, without predictive algorithms, indicating that a tractable, mapping-related component of functional dark matter and data sparsity in microbiome studies is directly addressable. Because Pfam- and BioCyc-derived mappings trade specificity for coverage, confidence in any individual reaction assignment depends on the supporting evidence tier and source database.

Humans↗

PubMed2026

RCoxNet: A Deep Learning Framework Integrating Random Walk with Restart, Mutation, and Clinical Data for Cancer Survival Prediction.

Accurate survival prediction in cancer remains challenging due to the sparsity of somatic mutation profiles and the failure of existing models to capture higher-order gene-gene dependencies. Network diffusion methods such as Random Walk with Restart (RWR) can propagate mutation signals across protein-protein interaction (PPI) networks to address sparsity, yet their integration within a deep learning Cox survival framework has not been comprehensively benchmarked across multiple cancer cohorts. We present RCoxNet, a deep learning framework that maps somatic mutation profiles onto a ConsensusPathDB-derived PPI network via RWR, selects prognostic genes by log-rank filtering, and processes network-informed mutation scores through three fully connected hidden layers feeding into a Cox proportional hazards output. RCoxNet was evaluated on The Cancer Genome Atlas (TCGA) cohorts for four cancer types (breast invasive carcinoma [BRCA], lung adenocarcinoma [LUNG], glioblastoma multiforme [GBM], and ovarian serous cystadenocarcinoma [OV]) using 20 independent random splits. The model achieved mean C-index values of 0.807 ± 0.044 (BRCA), 0.750 ± 0.039 (LUNG), 0.704 ± 0.041 (GBM), and 0.668 ± 0.036 (OV), consistently outperforming DeepSurv, Cox-nnet, SurvivalNet, Cox Elastic-Net (Cox-EN), and DeepHit, with statistically significant gains over Cox-EN, Cox-nnet, SurvivalNet, and DeepHit across the majority of cohorts. RCoxNet demonstrates that embedding sparse mutation profiles into a PPI network context substantially improves cancer survival prediction and yields biologically interpretable prognostic features relevant to precision oncology.

cancer survival prediction↗

PubMed2026

Benchmark of biomarker identification and prognostic modeling methods on diverse censored data.

The practices of identifying biomarkers and developing prognostic models using genomic data has become increasingly prevalent. Such data often features characteristics that make these practices difficult, namely high dimensionality, correlations between predictors, and sparsity. Many modern methods have been developed to address these problematic characteristics while performing feature selection and prognostic modeling, but a large-scale comparison of their performances in these tasks on diverse right-censored time to event data (aka survival time data) is much needed. We have compiled many existing methods, including some machine learning methods, several which have performed well in previous benchmarks, primarily for comparison in regards to variable selection capability, and secondarily for survival time prediction on many synthetic datasets with varying levels of sparsity, correlation between predictors, and signal strength of informative predictors. For illustration, we have also performed multiple analyses on a publicly available and widely used cancer cohort from The Cancer Genome Atlas using these methods. We evaluated the methods through extensive simulation studies in terms of the false discovery rate, F1-score, concordance index, Brier score, root mean square error, and computation time. Of the methods compared, CoxBoost and the Adaptive LASSO performed well in all metrics, and the LASSO and elastic net excelled when evaluating concordance index and F1-score. The Benjamini-Hoschberg and q-value procedures showed volatile performances in controlling the false discovery rate. Some methods' performances were greatly affected by differences in the data characteristics. With our extensive numerical study, we have identified the best performing methods for a plethora of data characteristics using informative metrics. This will help cancer researchers in choosing the best approach for their needs when working with genomic data.

Humans↗

PubMed2026

A divide and conquer strategy for recapitulating whole genome 3D structure using Hi-C data.

The three dimensional (3D) spatial organization of the genome is closely linked to biological functions and can be captured by Hi-C assays through interrogating genome-wide chromatin interactions. Methodologies for inferring 3D structures from Hi-C data summarized as a two-dimensional (2D) contact matrix can be broadly placed within the paradigms of optimization-based and sampling-based. Many optimization-based methods are capable of constructing whole genome 3D structures but do not account for spatial dependency in the 2D data matrix nor cell heterogeneity in bulk Hi-C data, which provide an average over millions of cells. Sampling-based methods, on the other hand, are probabilistic model-based and can account for not only dependency, heterogeneity, but also other features inherent in Hi-C data, such as over-dispersion and sparsity. However, whole-genome 3D structure recapitulation is too computationally expensive for sampling-based methods, while chromosome-by-chromosome strategies for sampling-based methods ignore important information on inter-chromosomal contacts. To address these issues, we propose the truncated Random effect EXpression-cut and paste (tREX-cap) method, which applies the tREX model within a divide and conquer strategy. The resulting method inherits the good data-feature-cognizant properties of tREX and, in the meantime, can efficiently infer the whole genome 3D structure. We demonstrate the performance of tREX-cap through an extensive simulation study and analyses of a Hi-C lymphoblastoid dataset and a Hi-C IMR90 dataset.

Humans↗

PubMed1977

Clear cell carcinoma of the ovary: light and electron microscopic studies.

Twelve cases of ovarian clear cell carcinoma were studied histologically. Four cases were examined electron microscopically and compared with other conditions. Five tumors were directly connected with ovarian endometriosis. They were histologically classified into a tubular type with hobnail cells and a solid type without tubular pattern or hobnail cells. Electron microscopic figures of the tumor cells are identical having large nuclei, abundant glycogen, lamellated RER, few lipid droplets, and sparse but well-developed microvilli. The basophilic dark cells frequently encountered in the tubular type are morphologically quite similar to clear cells excepting for sparsity of glycogen and lipid droplets. Alveolar arrangement of 6 to 10 tubular structures (honeycomb structure) resembling alveolate structure seen in late secretory endometrium was found in tumor cells of one case. Ultrastructural feature of clear cell carcinoma closely resemble those of Arias-Stella endometrium and clear cell carcinoma of endocervix suggesting their Muellerian origin.

Adenocarcinoma↗

PubMed1977

The fine structure of the melanocytes of the adult mouse hair follicle during their amelanotic phase (telogen and early anagen).

Melanocyte-precursor cells have been described previously in telogen and early anagen hair germs by observation of Epon sections using light microscopy. The existence of these precursor cells has not been generally accepted due to lack of ultrastructural evidence. It is our purpose to describe these cells using transmission electron microscopy. We studied hair germs of the dorsal skin of adult mice in the appropriate stages of the hair cycle prior to melanogenesis. Telogen melanocytes are distinguishable from adjacent keratocytes by their lack of desmosomes, lack of coarse clumps of microfilaments, sparsity of ribosomes, unusually few nuclear pores and presence of dendrites. Melanocytes in early anagen show extensive growth with abundant polysomes, elaborate RER and Golgi zones. They still differ from keratocytes in their lack of desmosomes and clumps of microfilament, and also in the occasional presence of premelanosomes.

Animals↗

PubMed1975

Transmission and scanning electron microscope studies of calcified cartilage resorption.

The authors' previous report (Savostin-Asling and Asling, '73) demonstrated that Meckel's carilage is a favorable site for study of calcified cartilage resorption. In the present study the ultrastructural features at this resorption front have been examined by transmission and scanning electron microscopes (19-day rat retus). Multinucleated giant cells chondroclasts) dominated the erosion front. The many features which they showed in common with osteoclasts included abundant mitochondria, vacuolation, lysonsomes, sparsity of rough-sufaced endoplasmic reticulum, and deep infoldings at loci of contact with calcified matrix. Crumbling of matrix (with mineral crystals penetrating between these foldings) and fragmentation of collagen fibrils were also seen. The propensity of chondroclasts for spanning several opened lacunae provided special opportunity to demonstrate cell surface modifications in presence or absence of matrix contact. Amebiod processes extending into lacunae were seen by both transmission and scanning procedures; they were sometimes tipped with a veil of filamentous processes as small as 0.3 mum in diameter. Most hypertrophic chondrocytes. when released from lacunae, appeared to be disintegrating. However, in accord with previous evidence of their possible merger with chondroclasts (in light microscopic studies) there was also evidence for breakdown of cell walls between a chondroclast and a chondrocyte in intimate contact, with possibility of cytoplasmic continuity.

Animals↗

PubMed1977

The thyroid gland of the woodchuck, Marmota monax: a morphological study of seasonal variations in the follicular cells.

The morphology of the thyroid gland of the woodchuck, Marmota monax, was studied during the four seasons of the year. In the spring the thyroid is extremely heterogenous in appearance. Some follicular cells appear quite active. They contain a well defined Golgi apparatus, abundant large colloid droplets and pseudopodia but few, if any, apical vesicles. Other less active cells have poorly defined rough surfaced endoplasmic reticulum and lack a well developed Golgi apparatus. They do not contain apical vesicles or colloid droplets. Summer thyroids have uniformly small follicles which are lined by high cuboidal cells containing numerous mitochondria, apical vesicles, abundant rough surfaced endoplasmic reticulum, and lipid droplets but few colloid droplets. There is extensive lateral and basal infolding of the cytoplasmic membranes in these cells. In the fall and winter the follicles are larger than in the summer and contain more colloid. Numerous heterogeneous dense bodies appear in the cytoplasm of the follicular cells in the fall and increase in number in the winter when there is an obvious sparsity of such glycoprotein synthetic organelles as Golgi apparatus and rough surfaced endoplasmic reticulum. These morphologic changes are compared with previous studies of thyroid structure and function in other animals and are correlated with the seasonal physiologic activities of the woodchuck.

Animals↗

PubMed2026

Quantile Tensor Regression for Integrative Genomic Analysis of Oesophageal Carcinoma.

Recent integrative genomic studies have increasingly exploited the tensor structure of multi-omics data to develop statistical methods that jointly model the relationship between clinical outcomes and multiple genomes. However, genomic measurements and clinical outcomes are frequently contaminated by outliers or heavy-tailed noise, necessitating robust tensor-based inference approaches. In this paper, we investigate the quantile tensor regression with an emphasis on the region selection problem. We introduce a novel estimator that integrates quantile regression for robustness with a nonconvex penalty to encourage sparsity in the tensor coefficient, thereby enabling the identification of localized genomic regions that significantly influence the clinical response. To solve the resulting optimization problem, we devise an effective algorithm tailored to the nonconvex objective and tensor architecture. We establish the asymptotic properties of the proposed nonconvex penalized estimator. Extensive simulations demonstrate the excellent finite-sample performance of the proposed estimator. We further illustrate the practical utility of the proposed estimator through an application to esophageal carcinoma data, providing empirical validation.

Humans↗

PubMed1975

Fine structural changes in the hamster pineal gland after blinding and superior cervical ganglionectomy.

Pineal glands of male hamsters 8 weeks after removal of both eyes or both superior cervical ganglia and those of untreated animals were studied by electron microscopy. In the blinded hamsters the reproductive organs were remarkably involuted, whereas the pinealocytes enlarged and were characterized by a tremendous hypertrophy of the smooth-surfaced endoplasmic reticulum, in the mesh of which some dense cored vesicles were distributed. In constrast the pinealocytes of ganglionectomized hamsters atrophied and were noted by a large number of lysosomes and sparsity of the agranular reticulum, the testes being significantly larger than the controls. The findings were interpreted to be compatible with the view that the pineal capability of producing antigonadal substance is augmented by blinding and is suppressed by cervical ganglionectomy due to the impairment of normal functioning of the pineal by denervation. It is assumed that the degree of development of the smooth endoplasmic reticulum in the pinealocyte is parallel to, and reflects that of the pineal activity. Probably the membranous organelle is involved in the elaboration of the pineal antigonadal substance.

Animals↗

PubMed1979

Various types of corticotectal neurons of cats as demonstrated by means of retrograde axonal transport of horseradish peroxidase.

The retrograde labeling of cortical neurons with horseradish peroxidase (HRP) was used to investigate the morphological features of neurons in various cortical areas projecting to the superior colliculus in the cat. Corticotectal cells were found to be labeled in layer V of the entire cerebral cortex. The number of labeled cells and their locations varied according to the sites of injections of HRP in the colliculus. Most of the corticotectal cells identified in the present study were small (9--20 micrometer in diameter, 66%) and medium (20--40 micrometer, 30%) pyramidal neurons and only 4% of them were large (more than 40 micrometer). The labeled cells, 261 in total number, had somal diameters of 20.8 +/- 8.0 micrometer (mean and SD). The range of sizes of the labeled neurons was different in different cortical areas. For example, the labeled neurons in the Clare-Bishop area had a greater proportion of large diameter cells than in other areas. The present findings are largely in agreement with the previous data of anterograde degeneration methods with respect to the topographical correlation of the corticotectal projections. However, in some cortical areas, e.g., the sensorimotor and the first visual (area 17) cortex of the lateral surface of the hemisphere, relatively small numbers of corticotectal neurons appear to have been labeled by retrogradely transported HRP. The sparsity of the labeled neurons in certain cortical areas may reflect the existence of corticotectal neurons with axon collaterals supplying brain structures other than the superior colliculus.

Animals↗

PubMed1979

Modulatory role for biogenic amines in the cerebral cortex. Microiontophoretic studies.

In order to investigate the mode of action of biogenic amines in rat cerebral cortex, the unitary activity of spontaneously firing neurons and their excitatory response to acetylcholine (ACh) were examined using microiontophoretic administration of dopamine (DA), noradrenaline (NA) and serotonin (5-HT). The predominant effect of these biogenic amines on the spontaneous activity was a profound and prolonged inhibition of firing (2-4 min), which attained its maximum within 15-120 sec. This response was generally more abrupt in onset and of greater magnitude with NA and 5-HT than with DA. Most units inhibited by DA, NA and 5-HT also showed marked depression of their excitatory response to ACh when pretreated with these biogenic amines. With repetitive administration of ACh, it could be shown that the total duration of inhibition of ACh responses by DA and NA was not as prolonged as the inhibition of the spontaneous firing of the same cells. With 5-HT, the initial ACh responses of many neurons could be completely blocked, and this inhibitory effect lasted as long as the inhibition of spontaneous firing. In view of the anatomical data demonstrating a relative sparsity of monoamine nerve terminals in cerebral cortex, the strong inhibition induced by DA, NA or 5-HT may have reflected slow inactivation of the biogenic amines. However, it could also be indicative of underlying mechanisms of action dependent on metabolic changes. Indeed, the interaction between biogenic amines and ACh might imply a balance between the intracellular pools of cAMP and cGMP is directly or indirectly influenced by the biogenic amines and ACh, respectively. This hypothesis would not exclude other modes of local interaction between DA, NA, 5-HT and ACh, and appears compatible with the modulatory role of biogenic amines in cerebral cortex.

Acetylcholine↗

PubMed1991

Apoptosis in the nervous system in experimental allergic encephalomyelitis.

We report here for the first time the occurrence of apoptosis of cells in the spinal cord in experimental allergic encephalomyelitis (EAE), an autoimmune, T-cell-mediated demyelinating disease. Four different forms of EAE were studied in the Lewis rat: (i) acute EAE induced by inoculation with whole spinal cord and adjuvants; (ii) acute EAE induced by inoculation with myelin basic protein (MBP) and adjuvants; (iii) acute EAE induced by the passive transfer of MBP-sensitized spleen cells; (iv) chronic relapsing EAE induced by inoculation with whole spinal cord and adjuvants followed by treatment with low-dose cyclosporin A. Cells undergoing apoptosis were recognized at light and electron microscopy by the presence of either crescentic masses of condensed chromatin lying against the nuclear envelope or rounded masses of uniformly dense chromatin. They were found in both the white and grey matter of the spinal cord in all 4 forms of this disease. Although it was not possible to identify definitively the types of cells undergoing apoptosis, the size and location of some of the affected cells suggested that they were oligodendrocytes. As there is now a large body of evidence that T-cell-induced target cell death takes the form of apoptosis, it is attractive to hypothesize that oligodendrocyte apoptosis is occurring in EAE as a result of oligodendrocyte-directed T-cell cytotoxicity. However, other apoptotic cells were located within the myelin sheath, meninges and perivascular spaces and were clearly not oligodendrocytes but were most likely blood-derived mononuclear cells. The sparsity of their cytoplasm and the absence of phagocytosed material suggested that they were mainly lymphocytes rather than macrophages. Apoptosis has been shown to be involved in deleting autoreactive T-cells during the normal development of tolerance. Thus apoptotic deletion of myelin/oligodendrocyte-specific lymphocytes in the central nervous system in EAE might explain both the subsidence of inflammation and the acquisition of tolerance in this autoimmune disease.

Animals↗

PubMed1991

Uranium content of blood, urine and hair of exposed and non-exposed persons determined by radiochemical neutron activation analysis, with emphasis on quality control.

Recent reviews have highlighted the diversity (and sparsity) of data for uranium concentrations in body fluids, bone and soft tissues, and it is unclear to what extent this reflects genuine geographical or biological variations. In the present work, a very sensitive radiochemical neutron activation analysis technique (detection limit, 1-2pg g-1) was applied to an exploratory study of uranium levels in hair, urine and blood of non-exposed and occupationally exposed persons. Since quality control in much previous work has been somewhat neglected, this aspect was emphasized by the use of a wide range of suitable reference materials, by standard addition experiments, and by care in sampling and sample handling. For controls, typical levels found in hair, urine and blood were 10 ng g-1, 10 and 5 ng l-1, respectively. The urine values correspond to the lower end of the literature data, while for blood our value is about two orders of magnitude lower than the presently quoted level. In the case of occupationally exposed persons, it was found that hair shows some promise as an indicator of exposure.

Hair↗