SEARCH · PubMed Health

Results for “Classification”

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 541 records · Page 30Linked to original sources

PubMed1999

Prognostic significance of the 1997 TNM classification of renal cell carcinoma.

PURPOSE: The TNM classification of renal cell carcinoma was recently revised in 1997. The most significant change from the previous edition (1987) is an increase in the size cutoff between T1 and T2 tumors from 2.5 to 7.0 cm. We compared the 1997 and 1987 TNM staging classifications in predicting patient outcome. MATERIALS AND METHODS: A total of 381 patients who underwent nephrectomy for renal cell carcinoma at our hospital between 1968 and 1994 were identified. Mean patient age was 61 years (range 15 to 89) and mean followup was 64.5 months. All pathological slides were re-reviewed in uniform manner and staged using the 1987 and 1997 TNM classifications. The impact of numerous pathological factors and each staging classification on disease specific survival and freedom from progression were statistically analyzed, and Kaplan-Meier survival curves were generated and compared. RESULTS: The 1997 TNM classification resulted in a redistribution of 170 cases previously classified as stage II (T2N0M0) to stage I (T1N0M0) under the new system. Both classifications were strong predictors of survival on univariate and multivariate analyses, and essentially equivalent in the ability to predict patient outcome. However, comparison of survival curves on Kaplan-Meier life tables revealed better separation of survival for stage I (T1N0M0) and stage II (T2N0M0) cases under the 1997 TNM classification, with survival for TNM stage I essentially remaining unchanged. CONCLUSIONS: The 1997 TNM classification of renal cell carcinoma appears to be equivalent to the previous classification in predicting outcome but permits better stratification of cases according to survival and, therefore, may have improved clinical usefulness.

Adolescent↗

PubMed1999

Linking the integrated management of childhood illness (IMCI) and health information system (HIS) classifications: issues and options.

Differences in the terms used to classify diseases in the Integrated Management of Childhood Illness (IMCI) guidelines and for health information system (HIS) disease surveillance could easily create confusion among health care workers. If the equivalent terms in the two classifications are not clear to health workers who are following the guidelines, they may have problems in performing the dual activities of case management and disease surveillance. These difficulties could adversely affect an individual's performance as well as the overall effectiveness of the IMCI strategy or HIS surveillance, or both. We interviewed key informants to determine the effect of these differences between the IMCI and HIS classifications on the countries that were implementing the IMCI guidelines. Four general approaches for addressing the problem were identified: translating the IMCI classifications into HIS classifications; changing the HIS list to include the IMCI classifications; using both the IMCI and HIS classification systems at the time of consultations; and doing nothing. No single approach can satisfy the needs of all countries. However, if the short-term or medium-term goal of IMCI planners is to find a solution that will reduce the problem for health workers and is also easy to implement, the approach most likely to succeed is translation of IMCI classifications into HIS classifications. Where feasible, a modification of the health information system to include the IMCI classifications may also be considered.

Case Management↗

PubMed2000

Tibial plateau fractures. A new classification scheme.

Fractures of the tibial plateaus are common injuries. Various classification schemes have been used to describe these injuries. Although each system has its own purpose, the simpler systems do not allow comparison with more complex divisions. The problem is compounded by the variable use of adjectives that describe these fractures. A comprehensive classification of tibial plateau fractures should group fractures that are similar in topography, morphology, and pathogenesis, requiring similar treatment, and having a similar prognosis. Fracture dislocations and standard tibial plateau fractures should be incorporated into a single classification to avoid the use of two complementary classifications. Any such classification should not be difficult to remember or to use. Keeping in mind these requirements, the authors devised a simple yet comprehensive classification. The authors studied 80 cases of tibial plateau fractures from January 1988 to September 1997, and used contemporary classifications of tibial plateau fractures as a database to formulate the new classification. A new fracture, subcondylar bicondylar with coronal split, has been classified for the first time. An alphanumeric system has been developed that has made nomenclature easy to remember and use. An effort has been made to address the profoundly confusing issue of variable adjectives that describe these injuries. A review of the literature shows that fractures in the authors' classification have been grouped according to similar pathomechanics, treatment, and functional results.

Adult↗

PubMed2000

Prognostic utility of the recently recommended histologic classification and revised TNM staging system of renal cell carcinoma: a Swiss experience with 588 tumors.

BACKGROUND: A new, internationally accepted histologic classification of renal cell carcinoma (RCC) and a new edition of the TNM staging system were introduced in 1997. In the latter, there was a dramatic change in the pT classification of organ-confined renal cancer in which the break point between category pT1 and pT2 was increased from 2.5 cm to 7 cm. METHODS: To study the significance of the new pT classification and the new recommendations for histologic classification, 588 nephrectomy specimens were reevaluated to define morphologic prognostic parameters in RCC. pT classification (TNM 1997), histologic subtype, histologic tumor grade, presence of necrosis, and sarcomatoid differentiation were assessed. RESULTS: The histopathologic review according to the new classification revealed 487 conventional (clear cell) (83%), 64 papillary (11%), 31 chromophobe (5%), and 6 collecting duct (1%) RCCs. Clinical follow-up was available for 470 RCCs. The new pT classification (1997) was strongly correlated with patient survival (P < 0.0001). Histologic grade, presence of necrosis, and sarcomatoid differentiation provided independent prognostic information on the clear cell subtype of renal cancer. Sarcomatoid differentiation, but not tumor necrosis, portended a dismal prognosis for patients with papillary RCC. Chromophobe RCC was associated with a significantly better prognosis than clear cell RCC (P = 0.05). Papillary RCC with scanty cytoplasm and small cells (type 1) behaved less aggressively than papillary tumors with eosinophilic cytoplasm and large cells (type 2; P < 0.001). CONCLUSIONS: Accurate histologic classification according to the new recommendations has implications because the prognostic importance of other histologic features that are of independent significance varies with tumor subtype. The data suggest that the new pT classification allows good separation of prognostic groups of patients with renal cancer.

Carcinoma, Renal Cell↗

PubMed2002

A new classification for crown fractures of teeth.

The aim of this work is to define a classification for traumatic injuries to crowns of teeth that offers a simplified interpretation, which can be communicated easily. These data will help us to evaluate the future restorations (composite resins, crowns, porcelain veneers) according to the variable design and extension of initial crown fracture lines. For this reason we conducted research on a vast sample of subjects, who presented different traumatic lesions, in order to obtain a specific selection of these typologies of injuries. We reviewed the literature for a complete, brief classification of simple use, which could serve our cause. The traditional classifications (Andreasen, Ellis, OMS) did not satisfy us for different reasons. In fact, they only consider the initial lesion situation, never focusing on the shape of the lesion and therefore never giving suggestions for the best kind of material for restoration. On the contrary, our interest is to define the kind of more valid materials according to the variable design of crown fracture lines, to foresee the duration of these materials and the best time to substitute them. To simplify and make our research a more affordable one, we created a 4 classes classification (A-B-C-D) and 3 subclasses (b1-c1-d1). Class A: all the simple enamel lesions which involve a mesial or distal crown angle, or only the incisal edge; Class B: all the enamel-dentin lesions, which involve a mesial or distal crown angle and the incisal edge. When a pulpal exposure exists we define it a Subclass b1; Class C: all the enamel-dentin lesions, which involve the incisal edge and at least a third of the crown surface. When a pulp exposure exists we define it a Subclass c1. Class D: all the enamel-dentin lesions, which involve a mesial or distal crown angle and the incisal or palatal surface, with root cement involvement (crown-root fractures). When a pulpal exposure exists we define it a Subclass d1. This classification showed how different kinds of lesions (Class B, b1, C and c1, in our new classification) fall under the same definition (enamel-dentinal fractures) in traditional classifications (Andreasen, Ellis). However, they need a wholly different clinical approach, and the material involved in the treatment shows different behavior and duration. This new classification simplified the gathering of data and the communication among practitioners, thus confirming its importance in getting optimal diagnostic and therapeutic protocols. It also allowed us to identify the most frequent crown fractures (Class B, b1 and C, c1) that in our sample of patients (age range: 8-18) are typically treated with composite resins or original fragment reattachment technique. All these studies brought us to develop this new "easy to use" classification of dental crown lesions that helped us to gather data easily, to choose the right materials to improve the communication among practitioners including by electronic means,

Adolescent↗

PubMed1971

A computerized English-Spanish correlation index to five biomedical library classification schemes based on MeSH.

A computerized English/Spanish correlation index to five biomedical library classification schemes and a computerized English/Spanish, Spanish/English listings of MeSH are described. The index was accomplished by supplying appropriate classification numbers of five classification schemes (National Library of Medicine; Library of Congress; Dewey Decimal; Cunningham; Boston Medical) to MeSH and a Spanish translation of MeSH The data were keypunched, merged on magnetic tape, and sorted in a computer alphabetically by English and Spanish subject headings and sequentially by classification number. SOME BENEFITS AND USES OF THE INDEX ARE: a complete index to classification schemes based on MeSH terms; a tool for conversion of classification numbers when reclassifying collections; a Spanish index and a crude Spanish translation of five classification schemes; a data base for future applications, e.g., automatic classification. Other classification schemes, such as the UDC, and translations of MeSH into other languages can be added.

Abstracting and Indexing↗

PubMedDate not supplied

[Evaluation of 3 functional classifications of cardiac insufficiency: a national multicenter study. National College of French Cardiologists].

The scope of this study is a comparison of three functional classifications: the New York Heart Association classification, the Duke University classification, and the specific activity scale we propose. The NYHA classification is subjective, difficult to reproduce and poorly correlated to the functional capacity measured ergometrically (duration of exercise and/or VO2). A new classification must therefore be proposed. A specific activity scale adapted to our national requirements appears to be the best solution. Our study covered 15 successive days. All patients suffering from congestive heart failure who were examined over this period were included. 700 liberal cardiologists took part in this study. 2353 patients were reviewed, mean age 69 (58% male, 42% female). Heart failure was of an ichaemic origin in 37% of cases, idiopathic in 25% and due to hypertension in 25%. Symptomatology was left sided in most cases. Among NYHA Class IV patients, 75% belonged to Class IV of the specific activity scale (SAS) (23% Class III, 1% Class II), and 88% of the Duke classification (10% Class III, 1% Class II). For NYHA Class III patients, 80% were SAS Class III (5% Class IV, 13% Class II), and only 38% (42% Class IV and 16% Class II) of the Duke classification. Regarding NYHA Class II patients, 74% were SAS Class II (21% Class III and 4% Class I), and 26% of the Duke classification (39% Class I, 29% Class III and 3.6% Class IV). Finally, among NYHA Class I patients, 60% were SAS Class I (34% Class II, 5% Class III), and 74% of the Duke classification (11% Class II and 13% Class III).(ABSTRACT TRUNCATED AT 250 WORDS)

Aged↗

PubMed1997

[TNM classification of carcinoma of the esophagus].

TNM classification of esophageal carcinoma was first described in the supplement to the first edition of the TNM classification in 1973. In the second edition, the classification was changed based on the data of 1,000 cases from the Task Force on Esophagus of American Joint Committee. In this edition, only the clinical classification was described, but the third edition included both clinical and post-surgical histopathological classification. But the criteria for T and pT classification differed. Before the fourth edition, specialists from Japan and the United States met in Hawaii in 1984. Data of the Japanese Nationwide Registration, including 7,742 patients from 1969 to 1978, were presented. After discussion based on these data, T was classified according to the depth of invasion, and perigastric lymph nodes were included in Regional Nodes in the fourth edition. Then, the TNM Research Committee of ISDE collected patient data of esophageal carcinoma from seven countries, and they were studied according to many factors. Based on these data, two proposals were made to the UICC TNM Committee. First, T1 should be divided into two categories: T1a, Tumor invasion of lamina propria; and T1b, Tumor invasion of submucosa. Second, metastases to distant lymph nodes should be grouped into the N classification instead of M classification. The first was accepted in the TNM Supplement of 1993, and the second will be accepted in the Fifth Edition, which will appear in 1997. It is important to accumulate data on many patients using the uniform registration form and to follow these patients very closely in the discussion of revisions to the TNM classification.

Esophageal Neoplasms↗

PubMed1998

The Novacode criteria for classification of ECG abnormalities and their clinically significant progression and regression.

Electrocardiographic (ECG) manifestations of clinical and subclinical cardiovascular disease are used as an important component in the evaluation of clinical trials, and there is an increasing demand for well-defined criteria for clinically significant evolution of ECG abnormalities. The Novacode ECG classification system provides a comprehensive hierarchical set of criteria for prevalent ECG abnormalities and for clinically significant serial ECG changes, both adverse and favorable, as a response to pharmacologic, surgical, and other interventions. These criteria are used to grade Q wave and ischemic abnormalities in order to achieve stable classification of both prevalent and incident myocardial infarctions by minimizing false classifications due to clinically insignificant ECG variations. This approach differs from the traditional Minnesota Code classification system, in which incident events are determined by changes in classification categories, with the application of additional elaborate validation rules to exclude frequent false classifications. Novacode hierarchy is so structured that for each abnormality, a general class is first determined with the simplest possible classification criteria and more specific abnormality subgroups are then classified with more elaborate criteria. This approach will satisfy differing needs of clinical trials for detail in classification. Explicit definition of ECG variables and condition statements for the classification criteria facilitate implementation of the Novacode with computer ECG programs.

Arrhythmias, Cardiac↗

PubMed2004

Classification issues related to neuropathic trigeminal pain.

The goal of a classification system of medical conditions is to facilitate accurate communication, to ensure that each condition is described uniformly and universally and that all data banks for the storage and retrieval of research and clinical data related to the conditions are consistent. Classification entails deciding which kinds of diagnostic entities should be recognized and how to order them in a meaningful way. Currently there are 3 major pain classification systems of relevance to orofacial pain: The International Association for the Study of Pain classification system, the International Headache Society classification system, and the Research Diagnostic Criteria for Temporomandibular Disorders (RDC/TMD). All use different methodologies, and only the RDC/TMD take into account social and psychologic factors in the classification of conditions. Classification systems need to be reliable, valid, comprehensive, generalizable, and flexible, and they need to be tested using consensus views of experts as well as the available literature. There is an urgent need for a robust classification system for neuropathic trigeminal pain.

Classification↗

PubMed2007

Development of an ecologic marine classification in the new zealand region.

We describe here the development of an ecosystem classification designed to underpin the conservation management of marine environments in the New Zealand region. The classification was defined using multivariate classification using explicit environmental layers chosen for their role in driving spatial variation in biologic patterns: depth, mean annual solar radiation, winter sea surface temperature, annual amplitude of sea surface temperature, spatial gradient of sea surface temperature, summer sea surface temperature anomaly, mean wave-induced orbital velocity at the seabed, tidal current velocity, and seabed slope. All variables were derived as gridded data layers at a resolution of 1 km. Variables were selected by assessing their degree of correlation with biologic distributions using separate data sets for demersal fish, benthic invertebrates, and chlorophyll-a. We developed a tuning procedure based on the Mantel test to refine the classification's discrimination of variation in biologic character. This was achieved by increasing the weighting of variables that play a dominant role and/or by transforming variables where this increased their correlation with biologic differences. We assessed the classification's ability to discriminate biologic variation using analysis of similarity. This indicated that the discrimination of biologic differences generally increased with increasing classification detail and varied for different taxonomic groups. Advantages of using a numeric approach compared with geographic-based (regionalisation) approaches include better representation of spatial patterns of variation and the ability to apply the classification at widely varying levels of detail. We expect this classification to provide a useful framework for a range of management applications, including providing frameworks for environmental monitoring and reporting and identifying representative areas for conservation.

Classification↗

PubMed1995

Classification trees for decision making in long-term care.

BACKGROUND: The purpose of the study was to develop a classification tool predicting a requirement for nursing home care in a population of nursing home applicants. In long-term care services, the objectives of classification mechanisms will include the prevention of inappropriate nursing home admission. METHOD: We studied 295 nursing home applicants residing in the Lower North Shore Area of Sydney, a high socioeconomic status area of Sydney, Australia. The predictor variables examined included: demographic data, social work assessment data, the presence of dementia and incontinence, the Barthel Index of Activities of Daily Living, and the Mini-Mental State Examination. RESULTS: Classification analysis using the C4.5 Program resulted in several classification trees for a decision for nursing home care with sensitivities greater than 70%. The best classification tree was one which combined the scores of the Barthel Index and the Mini-Mental State Examination. CONCLUSION: Classification trees in their simplicity of design and application have advantages over other analytical methods of classification. Classification analysis and the trees examined in this study may have future useful application in decision making for long-term care.

Classification↗

PubMed2003

Prognostic importance of physical examination for heart failure in non-ST-elevation acute coronary syndromes: the enduring value of Killip classification.

CONTEXT: In acute myocardial infarction, the presence and severity of heart failure at the time of initial presentation have been formally categorized by the Killip classification. Although well studied in ST-elevation myocardial infarction, the prognostic importance of Killip classification in non-ST-elevation acute coronary syndromes is not well established. OBJECTIVES: To determine the prognostic importance of physical examination for heart failure analyzed according to Killip classification in non-ST-elevation acute coronary syndromes and to understand its predictive value relative to other variables. DESIGN, SETTING, AND PATIENTS: From April 2001 to September 2003, We analyzed information from 26 090 patients with non-ST-elevation acute coronary syndromes enrolled in the GUSTO IIb, PURSUIT, PARAGON A, and PARAGON B trials. Demographic information was categorized by Killip class. Killip classes III and IV were combined into 1 category. Multivariate Cox proportional hazard models were developed to determine the prognostic importance of Killip classification in comparison with other variables. MAIN OUTCOME MEASURE: Association between Killip classification and all-cause mortality at 30 days and 6 months. RESULTS: Patients in Killip class II (n = 2513) and III/IV (n = 390) were older than those in Killip class I (n = 23 187), with higher rates of diabetes, prior myocardial infarction, ST depression, and elevated cardiac enzymes (all P<.001). Higher Killip class was associated with higher mortality at 30 days (2.8% in Killip class I vs 8.8% in class II vs 14.4% in class III/IV; P<.001) and 6 months (5.0% vs 14.7% vs 23.0%, respectively; P<.001). Patients with Killip class II, III, or IV constituted 11% of the overall population but accounted for approximately 30% of the deaths at both time points. In multivariate analysis, Killip class III/IV was the most powerful predictor of mortality at 30 days (hazard ratio [HR], 2.35; 95% confidence interval [CI], 1.69-3.26; P<.001) and 6 months (HR, 2.12; 95% CI, 1.63-2.75; P<.001). Killip class II was predictive of mortality at 30 days (HR, 1.73; 95% CI, 1.44-2.09; P<.001) and 6 months (HR, 1.52; 95% CI, 1.31-1.76; P<.001). Five factors-age, Killip classification, heart rate, systolic blood pressure, and ST depression-provided more than 70% of the prognostic information for 30-day and 6-month mortality. CONCLUSIONS: Killip classification is a powerful independent predictor of all-cause mortality in patients with non-ST-elevation acute coronary syndromes. Age, Killip classification, heart rate, systolic blood pressure, and ST depression should receive particular attention in the initial assessment of these patients.

Age Factors↗

PubMed1994

Lymph node classification systems in cutaneous T-cell lymphoma. Evidence for the utility of the Working Formulation of Non-Hodgkin's Lymphomas for Clinical Usage.

BACKGROUND: This study was undertaken to compare three classification schemes used to evaluate lymph nodes (LN) obtained from patients with cutaneous T-cell lymphoma (CTCL): a modified Rappaport classification, the National Cancer Institute-Veterans Administration (NCI-VA) classification based on the relative numbers of cerebriform cells in the paracortical areas, and the Dutch classification based on the presence of cerebriform cells with large nuclei in mycosis fungoides (MF) and diffuse infiltration by cerebriform cells in Sézary syndrome. METHODS: A study set of 195 LN obtained from patients with CTCL (MF, Sézary syndrome, and nonepidermotropic T-cell lymphomas) and 14 LN from patients with benign dermatoses was reviewed independently by three groups of pathologists familiar with each classification system. RESULTS: Each classification system provided useful prognostic information. However, contrary to prior reports, no significant difference in survival was apparent in patients with uneffaced LN when classified according to the NCI-VA (LN0-2 versus LN3) or Dutch (Gr0-1 versus Gr2) ratings. In addition, all classification systems demonstrated a poor survival time associated with effaced LN. By combining results from the modified Rappaport and Dutch classifications, three prognostic groups could be identified based on cell morphology: a low-grade category with a small cell histologic subtype (median survival time, 40 months); a high-grade immunoblastic subtype (median survival time, 9 months) composed of cells with an oval nucleus containing a large, usually solitary central nucleolus; and an intermediate-grade category composed of all cases without the distinctive small cell and immunoblastic morphologies (median survival time, 26 months). CONCLUSIONS: The authors propose that clearly involved LN in CTCL can be categorized on the basis of cell morphology into prognostic groups analogous to what has been proposed for the Working Formulation for Non-Hodgkin's Lymphomas for Clinical Usage.

Cell Transformation, Neoplastic↗

PubMed2002

The World Health Organization histologic classification system reflects the oncologic behavior of thymoma: a clinical study of 273 patients.

BACKGROUND: Although the histologic classification of thymic epithelial tumors has been confusing and controversial, an agreement on the universal classification system for thymic epithelial tumors was achieved by the World Health Organization (WHO) in 1999. The authors previously reported that the WHO histologic classification system reflects invasiveness and immunologic function of thymic epithelial tumors. In this subsequent study, they examined the prognostic significance of this classification system. METHODS: Clinical features as well as postoperative survival of patients with thymoma, but not thymic carcinoma, were examined with reference to WHO histologic classification based on an experience with 273 patients over a 44-year period. RESULTS: There were 18 type A tumors, 77 type AB tumors, 55 type B1 tumors, 97 type B2 tumors, and 26 type B3 tumors. In patients with type A, AB, B1, B2, and B3 tumors, the respective proportions of invasive tumor were 11.1%, 41.6%, 47.3%, 69.1%, and 84.6%; the respective proportions of tumors with involvement of the great vessels were 0%, 3.9%, 7.3%, 17.5%, and 19.2%; and the respective 20-year survival rates were 100%, 87%, 91%, 59%, and 36%. According to the Masaoka staging system, the 20-year survival rates were 89%, 91%, 49%, 0%, and 0% in patients with Stage I, II, III, IVa, and IVb disease, respectively. By multivariate analysis, the Masaoka staging system and the WHO histologic classification system were significant independent prognostic factors, whereas age, gender, association with myasthenia gravis, completeness of resection, or involvement of the great vessels were not significant independent prognostic factors. CONCLUSIONS: This study showed that histologic appearance reflects the oncologic behavior of thymoma when the WHO classification system is adopted. The WHO classification system may be helpful in clinical practice for the assessment and treatment of patients with thymoma.

Adolescent↗

PubMed2002

TNM residual tumor classification revisited.

BACKGROUND: For cancer patients, prognosis is strongly influenced by the completeness of tumor removal at the time of cancer-directed surgery or disease remission after nonsurgical treatment with curative intent. These parameters define the relative success of definitive treatment and can be codified by an additional subclassification within the TNM system, the residual tumor (R) classification. Despite the importance of residual tumor status in designing clinical management after treatment, misinterpretation and inconsistent application of the R classification frequently occur that diminish or abrogate its clinical utility. METHODS: An analysis of the relevant literature regarding the use and prognostic importance of the R classification was undertaken. RESULTS: In the current study, the prognostic importance of the R classification for different kinds of tumors is discussed. Problems that arise in using the R classification are described. Special issues regarding the use of the R classification are addressed. CONCLUSIONS: The R classification is a strong indicator of prognosis and facilitates the comparison of treatment results if applied in a consistent manner. Uniform use and interpretation of this classification is essential for the standardization of posttreatment data collection.

Humans↗

PubMed1983

Classification of immature and mature cells of the neutrophil series using morphometrical parameters.

Quantitative morphological data of six classes of immature and mature cells of the neutrophil series of the bone marrow of normal persons were used for statistical classification experiments (myeloblasts, promyelocytes, myelocytes, metamyelocytes, bands and segments). On each cell, parameters were measured directly from the image or calculated from the shape of the density histogram or the counting densitogram using a Texture Analysis System (E. Leitz, Wetzlar, Germany). The parameters were analyzed with the interactive statistical pattern recognition system ISPAHAN. One half of the data were used as a learning set and the other half as the test set. The parameters were compared according to their performance in discrimination between the classes, alone and in combinations. Parameters not contributing to an improvement of the discrimination were disregarded. Eleven parameters were selected and used for classification by two different methods: a stepwise and a "one-shot" method. Stepwise classification resulted in a 79% correct classification rate. Most errors occurred between cell classes in neighboring stages of maturation. In 96% of all cases the computer classification was either in accordance with that of the technician or with a cell class of a neighboring maturation stage. One step classification by the computer was in agreement with the technicians in 82% of the cases. For 98% of the cells the computer classification was either in accordance with that of the technician or with a cell class of a neighboring maturation stage. The data set was collected by two technicians, operating independently. Differences in their interpretation of the maturation stage were found by comparing the performance of classifiers based on both cell samples. Since the images of the cells were not available for reexamination, the causes of disagreement in classification between the technicians and between computer and technicians could not be evaluated.

Biometry↗

PubMed2004

Enzyme family classification by support vector machines.

One approach for facilitating protein function prediction is to classify proteins into functional families. Recent studies on the classification of G-protein coupled receptors and other proteins suggest that a statistical learning method, Support vector machines (SVM), may be potentially useful for protein classification into functional families. In this work, SVM is applied and tested on the classification of enzymes into functional families defined by the Enzyme Nomenclature Committee of IUBMB. SVM classification system for each family is trained from representative enzymes of that family and seed proteins of Pfam curated protein families. The classification accuracy for enzymes from 46 families and for non-enzymes is in the range of 50.0% to 95.7% and 79.0% to 100% respectively. The corresponding Matthews correlation coefficient is in the range of 54.1% to 96.1%. Moreover, 80.3% of the 8,291 correctly classified enzymes are uniquely classified into a specific enzyme family by using a scoring function, indicating that SVM may have certain level of unique prediction capability. Testing results also suggest that SVM in some cases is capable of classification of distantly related enzymes and homologous enzymes of different functions. Effort is being made to use a more comprehensive set of enzymes as training sets and to incorporate multi-class SVM classification systems to further enhance the unique prediction accuracy. Our results suggest the potential of SVM for enzyme family classification and for facilitating protein function prediction. Our software is accessible at http://jing.cz3.nus.edu.sg/cgi-bin/svmprot.cgi.

Amino Acid Sequence↗