Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “unsupervised clustering”

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 235 records · Page 13Linked to original sources
PubMed2005

Distinct molecular signature of inflammatory breast cancer by cDNA microarray analysis.

Inflammatory breast cancer (IBC) is a clinically distinct and aggressive form of locally advanced breast cancer with largely unknown genetic determinants. Overexpression of the RhoC GTPase and of HER2, and decreased ER-expression are involved in IBC. Multimodality treatment has increased survival but prognosis is still poor. Novel molecular targets for improved neoadjuvant treatment are necessary. Using cDNA microarrays, we performed genome-wide expression profiling of pre-treatment tumour samples of 16 patients with IBC and 18 patients with non-stage-matched non-IBC. Rigid clinical diagnostic criteria according to the TNM classification of the American Joint Committee on Cancer were adopted. Unsupervised hierarchical clustering accurately distinguished IBC and non-IBC samples. A set of 50 discriminator genes was identified in a learning group of tumour samples and was successful in diagnosing IBC in a validation group of samples (accuracy of 88%). Exclusion of ER-related or HER2-related genes did not alter this discriminatory accuracy, indicating that the expression of other genes in addition to ER and HER2 characterize the IBC phenotype. The molecular signature of IBC revealed the overexpression of a large number of NF-kappaB target genes, explaining at least part of the aggressive nature of IBC. Successful validation of some of the overexpressed genes by immunohistochemistry or real-time quantitative PCR demonstrated the robustness of the cDNA microarray experiments. The results of our study provide potential targets for the treatment of patients with IBC.

Belgium↗
PubMed2006

Intratumor genomic heterogeneity in breast cancer with clonal divergence between primary carcinomas and lymph node metastases.

Conflicting theories of epithelial carcinogenesis disagree on the clonal composition of primary tumors and on the time at which metastases occur. In order to study the spatial distribution of disparate clonal populations within breast carcinomas and the extent of the genetic relationship between primary tumors and regional metastases, we have analyzed by comparative genomic hybridization 122 tissue samples from altogether 60 breast cancer patients, including 34 tumor samples obtained from different quadrants of 9 breast carcinomas, as well as paired primary-metastatic samples from 12 patients. The median intratumor genetic heterogeneity score (HS) was 17.4% and unsupervised hierarchical clustering analysis comparing the genetic features to those of an independent series of 41 breast carcinomas confirmed intratumor clonal divergence in a high proportion of cases. The median HS between paired primary breast tumors and lymph node metastases was 33.3%, but the number of genomic imbalances did not differ significantly. Clustering analysis confirmed extensive clonal divergence between primary carcinomas and lymph node metastases in several cases. In the independent series of 41 breast carcinomas, the number of genomic imbalances in primary tumors was significantly higher in patients presenting lymph node metastases (median = 15.5) than in the group with no evidence of disease spreading at diagnosis (median = 5.0). We conclude that primary breast carcinomas may be composed of several genetically heterogeneous and spatially separated cell populations and that paired primary breast tumors and lymph node metastases often present divergent clonal evolution, indicating that metastases may occur relatively early during breast carcinogenesis.

Alleles↗
PubMed1990

Unsupervised learning and discriminant analysis applied to identification of high risk postoperative cardiac patients.

A set of 200 patients in the 6 hours immediately following cardiac surgery was analysed within a multidimensional space of 13 commonly monitored physiological variables in order to identify high risk patterns. The application of an unsupervised learning (clustering) method to these data clearly showed the existence of two well-separated classes of low and high risk patients. A stepwise discriminant analysis was then applied to patients representative of the two classes in order to find those variables which, over time, possessed the greatest separation power. The latter always included the oxygen delivery (DO2), an index related to the oxygen content in the blood (Pv(-)O2 or avO2D) and a myocardial contractility index (VF or LAP).

Algorithms↗
PubMed2004

Clinico-molecular study of dedifferentiation in well-differentiated liposarcoma.

Well-differentiated liposarcoma (WD) acquires fully malignant potential when the histological progression named dedifferentiation occurs. This progression is supposed to occur in a time-dependent manner but this is still a debated issue. Clinically, the prediction of dedifferentiation for WD is very important from the therapeutic point of view. To identify genes that are predictive of dedifferentiation and to understand the mechanism of dedifferentiation, we investigated clinical information of 50 cases and studied the gene expression profiles of 36 lipomatous tumors using cDNA microarray. The clinical study showed that the dedifferentiation did not always seem to occur in a time-dependent manner. Interestingly, from the gene expression study, unsupervised hierarchical clustering analysis of well-differentiated lesions obtained from dedifferentiated liposarcoma (DD) cases that were indistinguishable from WD pathologically showed a clearly distinct gene expression pattern from WD. Using the pattern-matching program, 1687 genes including 487 known genes were identified, which discriminated WD cases from well-differentiated lipomatous lesions obtained from DD cases. These results suggest that the dedifferentiation may arise from different types of WD that could be distinguished from gene expression profiling but could hardly be classified by the pathological studies.

Base Sequence↗
PubMed2006

Expression profilings of 39 genes selected by ANOVA could separate precursors of murine dendritic cells and macrophages.

Dendritic cells (DCs) and macrophages share some stages in the development and function of antigen presentation. But it is difficult to separate them from their precursors. We used one-way ANOVA (analysis of variances) on murine expression profilings of several hematopoietic cells associated with DCs and macrophages to find the genes with great differences across the cell groups. These groups were the DCs from spleen, cultivated DCs, DC precursors, DC progenitors, DC progenitor cell lines, hematopoietic stem cell (HSC), and bone marrow-derived macrophages. The data of expression profilings were all downloaded from GEO and ArrayExpress database. After the normalization of 11 housekeeping genes across 42 arrays, we got 39 genes (44 probesets) by analysis of one-way ANOVA (Bonferroni step-down) with p values cutoff of 0.05. These genes (probesets) could separate the hematopoietic cells well by the methods of unsupervised hierarchical clustering and principal component analysis (PCA). The class prediction also indicated that these genes could separate the precursors of DC and macrophages with 20 arrays composed of 5 cell types with the same normalization. The accuracy rate of class prediction was 90% (18/20). The genes selected by one-way ANOVA included those of MHC (major histocompatibility complex) and defense of immunity, cell adhesion, chemokine or its receptors, and transcription factors. The results indicated that these 39 genes could separate precursors of DC and macrophages very clearly. It was suggested that these genes might represent some important molecules that related with the precursors of DCs and macrophages, and were worthy for further study.

Animals↗
PubMed2024

Quantitative natural history modeling of HPDL-related disease based on cross-sectional data reveals genotype-phenotype correlations.

PURPOSE: Biallelic HPDL variants have been identified as the cause of a progressive childhood-onset movement disorder, with a broad clinical spectrum from severe neurodevelopmental disorder to juvenile-onset pure hereditary spastic paraplegia type 83. This study aims at delineating the geno- and phenotypic spectra of patients with HPDL-related disease, quantitatively modeling the natural history, and uncovering genotype-phenotype associations. METHODS: A cross-sectional analysis of 90 published and 1 novel case was performed, using a Human-Phenotype-Ontology-based approach. Unsupervised phenotypic clustering was used alongside in silico analyses to identify distinct patient subgroups. RESULTS: The study models the natural history of the HPDL-related disease in a global cohort, clarifying the molecular and phenotypic spectrum and identifying 3 distinct subgroups characterized by differences in onset, clinical trajectories, and survival. It establishes genotype-phenotype associations, showing that the presence of moderately pathogenic missense variants in 1 allele leads to a milder, spastic paraplegic phenotype with later disease onset, whereas biallelic, highly pathogenic missense or truncating variants are associated with a more severe phenotype and reduced life span. CONCLUSION: Quantitative and unbiased natural history modeling in HPDL-related disease reveals significant genotype-phenotype associations, providing a foundation for variant interpretation, anticipatory guidance, and choice of outcome measures in future prospective and functional studies.

Humans↗
PubMed2004

Identification of carboxypeptidase E and gamma-glutamyl hydrolase as biomarkers for pulmonary neuroendocrine tumors by cDNA microarray.

Pulmonary neuroendocrine tumors vary dramatically in their malignant behavior. Their classification, based on histological examination, is often difficult. In search of molecular and prognostic markers for these tumors, we used cDNA microarray analysis of human transcripts against reference RNA from a well-characterized immortalized bronchial epithelial cell line, BEAS-2B. Tumor cells were isolated by laser-capture microdissection from primary tumors of 17 typical carcinoids, small cell lung cancers, and large cell neuroendocrine carcinomas. An unsupervised, hierarchical clustering algorithm resulted in a precise classification of each tumor subtype according to the proposed histological classification. Selection of genes, using supervised analysis, resulted in the identification of 198 statistically significant genes (P <.004) that also accurately discriminated between 3 predefined tumor subtypes. Two-by-two comparisons of these genes identified classifier genes that distinguished each tumor subtype from the others. Changes in expression of selected differentially expressed genes for each tumor subtype were internally validated by real-time reverse-transcription polymerase chain reaction. Expression of 2 potential classifier gene products, carboxypeptidase E (CPE) and gamma-glutamyl hydrolase (GGH), was validated by immunohistochemistry and cross-validated on additional archival samples of pulmonary neuroendocrine tumors. Kaplan-Meier survival analysis revealed that immunostaining for CPE was a statistically significant predictor of good prognosis, whereas GGH expression correlated with poor prognosis. Thus, cDNA microarray analysis led to the identification of 2 novel biomarkers that should facilitate molecular diagnosis and further study of pulmonary neuroendocrine tumors.

Biomarkers, Tumor↗
PubMed2005

Differential gene expression in anaplastic lymphoma kinase-positive and anaplastic lymphoma kinase-negative anaplastic large cell lymphomas.

Anaplastic large cell lymphoma (ALCL) is an aggressive large T- or null-cell lymphoma. Most ALCLs arising in children and young adults express a constitutively active receptor tyrosine kinase, anaplastic lymphoma kinase (ALK). Anaplastic large cell lymphomas lacking ALK are clinically heterogeneous and their pathogenesis is unknown. This study is the first complementary DNA (cDNA) microarray analysis using RNA extracted from tumor tissue (7 ALK+ ALCLs and 7 ALK- ALCLs) to identify genes differentially expressed or shared between the ALK+ and ALK- tumors. Unsupervised hierarchical clustering using the top 11 most statistically significant discriminator cDNAs correctly grouped all ALK+ and ALK- tumors. Hierarchical clustering analysis using the 44 cDNAs with the greatest differential expression between ALK+ and ALK- RNAs grouped 6 of 7 ALK+ ALCLs together and 1 ALK+ ALCL with the ALK- group. In general, ALK+ tumors overexpress genes encoding signal transduction molecules (SYK , LYN , CDC37) and underexpress transcription factor genes (including HOXC6 and HOX A3 ) compared with the ALK- group. Cyclin D3 was overexpressed in the ALK+ group and the cell cycle inhibitor p19INK4D was decreased in the ALK- group, suggesting different mechanisms of promoting G 1 /S transition. Both groups had similar proliferation rates. Genes highly expressed in both ALK- and ALK+ ALCLs included kinases (LCK, protein kinase C, vav2, and NKIAMRE) and antiapoptotic molecules, suggesting possible common pathogenetic mechanisms as well.

Adolescent↗
PubMed2006

Application of transcriptional and biological network analyses in mouse germ-cell transcriptomes.

Serial analysis of gene expression (SAGE) provides a global analysis platform for profiling mRNA populations present in cells of interest without the constraint of gene selection and the ambiguous nature of data obtained. However, most of the reports on SAGE and germ cell development are limited to descriptive analyses. Here, we report a series of bioinformatic analyses using recently published SAGE data on the transcriptome of mouse type A spermatogonia (Spga), pachytene spermatocytes (Spcy), and round spermatids (Sptd). Tags with a total count of > or =20 in three SAGE libraries were examined. Our aim was to identify and discover potential transcriptional regulators and pathways involved at different stages of spermatogenesis. Unsupervised hierarchical clustering based on tag expression and Gene Ontology analysis were applied to identify genes and biological processes overrepresented at a particular stage of development. The 5' cis-regulatory elements were examined for common regulators in different functional clusters. Potential biological networks were also constructed to reveal the link between the gene candidates. Biological pathways related to the three germ cell stages were constructed. A number of known transcription regulators in spermatogenesis, including NF-kappaB, SP1, AP-1, and EGR, were identified. Novel promoter elements such as the E box in Spga-specific genes, GATA in Spcy-specific genes, and GKLF in Sptd-specific genes were also observed. Taken together, our approach is reliable and provides a foundation for the generation of novel biological hypotheses for studying spermatogenesis.

Animals↗
PubMed2007

Identification of genes with correlated patterns of variations in DNA copy number and gene expression level in gastric cancer.

To identify DNA copy number changes that had a direct influence on mRNA expression in gastric cancer, cDNA microarray-based comparative genomic hybridization (aCGH) and gene expression profiling were performed using 17 K cDNA microarrays. A set of 158 genes showing Pearson correlation coefficients over 0.6 between DNA copy number changes and mRNA expression level variations was selected. In an independent gene expression profiling of 60 tissue samples, the 158 genes were able to distinguish most of the normal and tumor tissues in an unsupervised hierarchical clustering, suggesting that the differential expression patterns displayed by this specific group of genes are most likely based on the gene copy number changes. Furthermore, 43 statistically significant (P<0.01) genes were selected that correctly distinguished all of the tissue samples. The copy number changes detected by aCGH can be verified by fluorescence in situ hybridization and real-time polymerase chain reaction. The selected genes include those that were previously identified as being tumor suppressors or deleted in various tumors, including GATA binding protein 4 (GATA4), monoamine oxidase A (MAOA), cyclin C (CCNC), and oncogenes including malignant fibrous histiocytoma amplified sequence 1 (MFHAS1/MASL1), high mobility group AT-hook 2 (HMGA2), PPAR binding protein (PPARBP), growth factor receptor-bound protein 7 (GRB7), and TBC1 (tre-2, BUB2, cdc16) domain family, member 1 (TBC1D1).

Cluster Analysis↗
PubMed2004

Gene expression profiling of tamoxifen-associated uterine cancers: evidence for two molecular classes of endometrial carcinoma.

OBJECTIVE: Endometrial cancers associated with tamoxifen exposure potentially represent a unique resource for investigating the molecular mechanisms of estrogen-induced tumorigenesis. The purpose of this study was to test the hypothesis that tamoxifen-associated endometrial carcinomas have a distinct gene expression profile compared to matched cases not associated with this exposure. METHODS: Microarray-based expression profiling was performed on a series of tamoxifen-associated (N = 10) and matched sporadic cases (N = 29) of endometrial carcinoma. RESULTS: Supervised class comparison revealed no statistically significant difference between the two groups (P = 0.48). However, unsupervised hierarchical clustering of the entire sample of tumors revealed two groups with extremely diverse molecular profiles that were independent of tamoxifen exposure. Of recognized clinicopathological factors, histologic grade correlated best with these two molecular classes. CONCLUSIONS: These data suggest that the molecular profile of tamoxifen-associated endometrial cancers is not different from that of endometrial carcinomas generally, and further, that there exist two highly distinct molecular subtypes of endometrial carcinoma.

Aged↗
PubMed2005

Distinction between serous tumors of low malignant potential and serous carcinomas based on global mRNA expression profiling.

OBJECTIVES: The molecular pathogenesis of ovarian serous tumors of low malignant potential (S-LMP) is not well understood, although the collective data suggest that they arise through molecular mechanisms distinct from those leading to conventional serous carcinomas (S-Ca). To further examine the molecular differences between these two diseases, we studied the gene expression pattern of ovarian S-LMP and S-Ca using high-density spotted cDNA and tissue microarrays. METHODS: Total RNA from 23 ovarian S-LMP and S-Ca was analyzed on 43,200 spot cDNA microarrays and the differential expression of proteins encoded by differentially expressed genes was validated using tissue microarrays. RESULTS: Unsupervised hierarchical clustering analysis of filtered data showed a complete separation between S-LMP and S-Ca, based predominantly on a small set of genes expressed at higher levels in S-LMP than in S-Ca. Many genes previously identified as up-regulated in ovarian carcinoma relative to normal ovarian tissue were expressed at even higher levels in S-LMP. These genes included mucin-1, mesothelin, HE4, PAX 8, and apolipoprotein J/clusterin. Immunohistochemical staining of tissue microarrays confirmed higher expression of selected proteins encoded by these genes in the S-LMP. Few genes were expressed at a higher level in S-Ca; these included E2F1, topoisomerase IIalpha, and cyclin E, with higher levels of cyclin E protein confirmed by immunohistochemistry. CONCLUSIONS: S-LMP and S-Ca are distinguished at the molecular level by a relatively small gene set, suggesting the pathogenesis of S-LMP as well as S-Ca may involve molecular pathways that escape detection by global gene expression profiling. In order to obtain biologically and clinically relevant information about the mechanisms involved in ovarian carcinogenesis, future studies based on molecular profiles of ovarian cancer should include analyses of low malignant potential tumors. Inclusion of such tumors is also critical to the evaluation of the efficacy of potential new diagnostic and/or therapeutic biomarkers.

Adult↗
PubMed2003

Distinct transcriptional profiles of adrenocortical tumors uncovered by DNA microarray analysis.

Comprehensive expression profiling of tumors using DNA microarrays has been used recently for molecular classification and biomarker discovery, as well as a tool to identify and investigate genes involved in tumorigenesis. Application of this approach to a cohort of benign and malignant adrenocortical tissues would be potentially informative in all of these aspects. In this study, we generated transcriptional profiles of 11 adrenocortical carcinomas (ACCs), 4 adrenocortical adenomas (ACAs), 3 normal adrenal cortices (NCs), and 1 macronodular hyperplasia (MNH) using Affymetrix HG_U95Av2 oligonucleotide arrays representing approximately 10,500 unique genes. The expression data set was used for unsupervised hierarchical cluster analysis as well as principal component analysis to visually represent the expression data. An analysis of variance on the three classes (NC, ACA plus MNH, and ACC) revealed 91 genes that displayed at least threefold differential expression between the ACC cohort and both the NC and ACA cohorts at a significance level of P < 0.01. Included in these 91 genes were those known to be up-regulated in adrenocortical tumors, such as insulin-like growth factor (IGF2), as well as novel differentially expressed genes such as osteopontin (SPP) and serine threonine kinase 15 (STK15). Increased expression of IGF2 was identified in 10 of 11 ACCs (90.9%) and was verified by quantitative reverse transcriptase-polymerase chain reaction. Select proliferation-related genes (TOP2A and Ki-67) were validated at the protein level using immunohistochemistry and adrenocortical tissue microarrays. Our results demonstrated significant and consistent gene expression changes in ACCs compared to benign adrenocortical lesions. Moreover, we identified several genes that represent potential diagnostic markers and may play a role in the pathogenesis of ACC.

Adenoma↗
PubMed2002

Gene expression profiling of cell lines derived from T-cell malignancies.

The expression profiles of eight cell lines derived from T-cell malignancies were compared to CD4-positive T-cells using cDNA microarray technology. Unsupervised hierarchical clustering of 4364 genes demonstrated substantial heterogeneity resulting in four distinct groups. While no genes were found to be uniformly up- or down-regulated across all cell lines, we observed 111 over-expressed genes (greater than two-fold) and 1118 down-regulated genes (greater than two-fold) in the lymphomas as a group when compared to CD4-positive T-cells. These included genes involved in cytokine signaling, cell adhesion, cytoskeletal elements, nuclear transcription factors, and known oncogenes and tumor suppressor genes. Quantitative fluorescent reverse transcription-polymerase chain reaction analysis demonstrated 70% concordance with the microarray results. While freshly isolated malignant cells may differ in their individual expression patterns relative to established cell lines from the same diagnoses, we feel that the variety of different lymphocytic cell lines that we examined provides a representative picture of the molecular pathogenesis of T-cell malignancies.

CD4-Positive T-Lymphocytes↗
PubMed2001

A dynamical model for the analysis and acceleration of learning in feedforward networks.

A dynamical system model is derived for feedforward neural networks with one layer of hidden nodes. The model is valid in the vicinity of flat minima of the cost function that rise due to the formation of clusters of redundant hidden nodes with nearly identical outputs. The derivation is carried out for networks with an arbitrary number of hidden and output nodes and is, therefore, a generalization of previous work valid for networks with only two hidden nodes and one output node. The Jacobian matrix of the system is obtained, whose eigenvalues characterize the evolution of learning. Flat minima correspond to critical points of the phase plane trajectories and the bifurcation of the eigenvalues signifies their abandonment. Following the derivation of the dynamical model, we show that identification of the hidden nodes clusters using unsupervised learning techniques enables the application of a constrained application (Dynamically Constrained Back Propagation-DCBP) whose purpose is to facilitate prompt bifurcation of the eigenvalues of the Jacobian matrix and, thus, accelerate learning. DCBP is applied to standard benchmark tasks either autonomously or as an aid to other standard learning algorithms in the vicinity of flat minima. Its application leads to significant reduction in the number of required epochs for convergence.

Acceleration↗
PubMedDate not supplied

Automatic segmentation of non-enhancing brain tumors in magnetic resonance images.

Tumor segmentation from magnetic resonance (MR) images may aid in tumor treatment by tracking the progress of tumor growth and/or shrinkage. In this paper we present the first automatic segmentation method which separates non-enhancing brain tumors from healthy tissues in MR images to aid in the task of tracking tumor size over time. The MR feature images used for the segmentation consist of three weighted images (T1, T2 and proton density (PD)) for each axial slice through the head. An initial segmentation is computed using an unsupervised fuzzy clustering algorithm. Then, integrated domain knowledge and image processing techniques contribute to the final tumor segmentation. They are applied under the control of a knowledge-based system. The system knowledge was acquired by training on two patient volumes (14 images). Testing has shown successful tumor segmentations on four patient volumes (31 images). Our results show that we detected all six non-enhancing brain tumors, located tumor tissue in 35 of the 36 ground truth (radiologist labeled) slices containing tumor and successfully separated tumor regions from physically connected CSF regions in all the nine slices. Quantitative measurements are promising as correspondence ratios between ground truth and segmented tumor regions ranged between 0.368 and 0.871 per volume, with percent match ranging between 0.530 and 0.909 per volume.

Astrocytoma↗
PubMed2005

Molecular classification of renal tumors by gene expression profiling.

Renal tumor classification is important because histopathological subtypes are associated with distinct clinical behavior. However, diagnosis is difficult because tumor subtypes have overlapping microscopic characteristics. Therefore, ancillary methods are needed to optimize classification. We used oligonucleotide microarrays to analyze 31 adult renal tumors, including clear cell renal cell carcinoma (RCC), papillary RCC, chromophobe RCC, oncocytoma, and angiomyolipoma. Expression profiles correlated with histopathology; unsupervised algorithms clustered 30 of 31 tumors according to appropriate diagnostic subtypes while supervised analyses identified significant, subtype-specific expression markers. Clear cell RCC overexpressed proximal nephron, angiogenic, and immune response genes, chromophobe RCC oncocytoma overexpressed distal nephron and oxidative phosphorylation genes, papillary RCC overexpressed serine protease inhibitors, and extracellular matrix products, and angiomyolipoma overexpressed muscle developmental, lipid biosynthetic, melanocytic, and distinct angiogenic factors. Quantitative reverse transcriptase-polymerase chain reaction and immunohistochemistry of formalin-fixed renal tumors confirmed overexpression of proximal nephron markers (megalin/low-density lipoprotein-related protein 2, alpha-methylacyl CoA racemase) in clear cell and papillary RCC and distal nephron markers (beta-defensin 1, claudin 7) in chromophobe RCC/oncocytoma. In summary, renal tumor subtypes were classified by distinct gene expression profiles, illustrating tumor pathobiology and translating into novel molecular bioassays using fixed tissue.

Gene Expression Profiling↗
PubMed2004

In situ spectral monitoring of mRNA translation in embryonic stem cells during differentiation in vitro.

Raman microspectroscopy was used to determine biochemical markers during the differentiation of embryonic murine stem cells (mES) in vitro. Such markers are useful to determine the differentiation status of ES cells cultured on biomaterials. Raman spectra of mES cells as undifferentiated, spontaneously differentiated (4 days), and differentiated cells via formation of embryoid bodies (16, 20 days) were analyzed. Unsupervised hierarchical cluster analysis and principal component analysis were used to determine biochemical differences between mES cells in various states of differentiation. The undifferentiated cells were characterized by high scores of the first principal component (PC1, 49% variance). Similarity between the PC1 loading and the Raman spectrum of RNA indicated a high concentration of RNA in mES cells compared to differentiated cells. The ratio between the peak areas of RNA and proteins was used as a measure of mRNA translation. Using the same peak area ratio, it was possible to differentiate even between mES as undifferentiated and in early stages of differentiation (4 days). These findings were correlated with biological studies reporting high levels of nontranslated mRNA during early embryonic development. Therefore, the RNA translation obtained from the Raman spectra can be used as marker of differentiation state of mES cells.

Animals↗