Federico Stefanini - Academia.edu (original) (raw)

Papers by Federico Stefanini

International Conference on Networking, Sensing and Control, 1999

Most of the approaches developed in the literature to elicit the a-priori distribution on Directe... more Most of the approaches developed in the literature to elicit the a-priori distribution on Directed Acyclic Graphs (DAGs) require a full specification of graphs. Nevertheless, expert's prior knowledge about conditional independence relations may be weak, making the elicitation task troublesome. Moreover, the detailed specification of prior distributions for structural learning is NP-Hard, (7), making the elicitation of large net- works

In Vitro Cellular & Developmental Biology - Animal

Bayesian Networks are increasingly used to represent conditional in- dependence relations among v... more Bayesian Networks are increasingly used to represent conditional in- dependence relations among variables and causal information in problem domains in which decisions are based on probabilistic reasoning. Structural learning is NP- hard therefore the database of observed cases must be often supplemented with search heuristics based on prior information. In this paper we present a software package for R, called

Alternatives to laboratory animals : ATLA, 2011

The human carcinogenicity evaluation of chemicals has a great impact on public health. In vitro m... more The human carcinogenicity evaluation of chemicals has a great impact on public health. In vitro methods, such as the cell transformation assay (CTA), allow for a fast and reliable assessment of the carcinogenic potential of a chemical compound in comparison with the standard two-year bioassay. The scoring and classification of foci in selected cell lines is performed, after staining, by light microscopy. Foci can be separated into three classes: type I, which are scored as non-transformed, and types II and III that are considered to include fully transformed foci. However, in a number of cases, even an expert is uncertain about the attribution of a focus to a given class, due to its mixed or intermediate nature. Here, we suggest a simple approach to classifying mixed or intermediate foci by exploiting the quantitative information available from images, which is captured by statistical descriptors. A quantitative index is proposed, to describe the degree of dissimilarity of mixed and...

Imaging, Manipulation, and Analysis of Biomolecules, Cells, and Tissues VIII, 2010

The Cell Transformation Assay (CTA) is one of the promising in vitro methods used to predict huma... more The Cell Transformation Assay (CTA) is one of the promising in vitro methods used to predict human carcinogenicity. The neoplastic phenotype is monitored in suitable cells by the formation of foci and observed by light microscopy after staining. Foci exhibit three types of morphological alterations: Type I, characterized by partially transformed cells, and Types II and III considered to have

Lecture Notes in Computer Science, 2007

Most of the approaches developed in the literature to elicit the a-priori distribution on Directe... more Most of the approaches developed in the literature to elicit the a-priori distribution on Directed Acyclic Graphs (DAGs) require a full specification of graphs. Nevertheless, expert’s prior knowledge about conditional independence relations may be weak, making the elicitation task troublesome. Moreover, the detailed specification of prior distributions for structural learning is NP-Hard, making the elicitation of large networks impractical. This

Biometrical Letters, 2013

Journal of the Royal Statistical Society: Series A (Statistics in Society), 2003

The evaluation of nuclear DNA evidence for identification purposes is performed here taking accou... more The evaluation of nuclear DNA evidence for identification purposes is performed here taking account of the uncertainty about population parameters. Graphical models are used to detail the hypotheses being debated in a trial with the aim of obtaining a directed acyclic graph. Graphs also clarify the set of evidence that contributes to population inferences and they also describe the conditional independence structure of DNA evidence. Numerical illustrations are provided by re-examining three case-studies taken from the literature. Our calculations of the weight of evidence differ from those given by the authors of case-studies in that they reveal more conservative values.

Journal of Microscopy, 2009

Carcinogenesis is a multi-step process involving genetic alterations and non-genotoxic mechanisms... more Carcinogenesis is a multi-step process involving genetic alterations and non-genotoxic mechanisms. The in vitro cell transformation assay allows the monitoring of the neoplastic phenotype by foci formation in suitable cells (e.g. C3H10T1/2 mouse embryo fibroblasts) showing aberrant morphology of massive build-up, polar and multi-layered densely stained cells. The classification of transformed foci in C3H cells relies on light microscopy scoring by a trained human expert based on standard rules. This procedure is time-consuming and prone, in some cases, to subjectivity, thereby leading to possible over- or under-estimation of the carcinogenic potential of tested compounds. Herewith we describe the in vitro neoplastic transformation induced by B[a]P and CdCl2, and the development of a foci classifier based on image analysis and statistical classification. The image analysis system, which relies on 'spectrum enhancement', is quantitative and extracts descriptors of foci texture and structure. The statistical classification method is based on the Random Forest algorithm. We obtained a classifier trained by using expert's supervision with a 20% classification error. The proposed method could serve as a basis to automate the in vitro cell transformation assay.

Journal of Applied Microbiology, 2007

To assess the effect of different concentrations of gossypol (0, 2, 4, 10 and 20 mg l(-1)) in com... more To assess the effect of different concentrations of gossypol (0, 2, 4, 10 and 20 mg l(-1)) in combination with NaCl and Na(2)SO(4) (20 mS cm(-1)) on the conidial germination and viability of Fusarium oxysporum f.sp. vasinfectum (Fov). A multinomial logistic model was developed to estimate the germination probability of Fov. The inhibitory effect was markedly evident at the two highest concentrations of gossypol; it varied among the isolates tested and with time, and it was attenuated by the presence of sodium salts. The inhibition was temporary as the germination probability increased after 8 h. Fluorescent staining revealed that gossypol either killed the conidia or retarded the elongation of the germ tubes. Fov showed the ability to overcome gossypol inhibition over time, and the inhibitory effect is reduced under saline conditions. Differential responses among Fov isolates to the presence of gossypol suggest that gossypol tolerance is genetically determined in the pathogen. This study suggests that selecting for high plant gossypol cultivars would have minimal effect on the overall Fov resistance of cotton. A new statistical model was developed to explore the statistical significance of plant-pathogen interactions.

Euphytica, 1995

Potato plantlets derived from in vitro propagation of three cultivars known for their field resis... more Potato plantlets derived from in vitro propagation of three cultivars known for their field resistance (Cruza 148 and BR-63.65) or susceptibility (Drsirre) to Pseudomonas solanacearum E.E Smith were artificially inoculated under controlled conditions. The aim of this work was to determine the optimal inoculum concentration and the best observation period in which the cultivars would show different reactions to bacterial infection as expected on the basis of their field performance.

Canadian Journal of Forest Research, 1995

To determine whether cypress bark canker, caused by the fungus Seiridium cardinale (Wag.) Sutton ... more To determine whether cypress bark canker, caused by the fungus Seiridium cardinale (Wag.) Sutton & Gibson, affects height and diameter growth of common cypress (Cupressus sempervirens L.), these parameters were measured over 4 years on inoculated and uninoculated ramets of 50 cypress clones at two locations in Italy. An analysis of covariance, using the pre-trial diameter and height of the ramets as the covariates, showed that the disease had no effect on either height or diameter growth.

Bioinformatics, 1993

Problem-solving and modelling within a biological context often need a level of descriptive accur... more Problem-solving and modelling within a biological context often need a level of descriptive accuracy that is unlikely to be capable of analytical treatment, especially if the mathematical background of the biologist is poor. Furthermore solver-model maintenance is often difficult without the availability of trained specialists. Better prospects are found in the genetic algorithm field. Genetic algorithms are a set of procedures formulated to solve complex problems without specifying rules for intermediate steps. This approach becomes feasible performing a Monte Carlo simulation of the natural evolution process, in which population improvement (search for solutions) in a considered environment (the specific problem domain) is achieved by following the genetic paradigm. Starting with a randomly constituted sample of individuals, drawn from the population of admissible values and expressed as binary strings, random mating brings about individuals of the next generation. Parents are chosen with a greater probability as the number of constraints violated by each individual becomes smaller. During the constitution of each generation the presence of some genetic operators causes the improvement of population diversity and its maintenance. Genetic operators are simple string transformation rules, generally independent of a specific context. We have developed the constant core of a minimal genetic algorithm, from which can be derived genetic problem-solvers in specific domains. An applicative example--a constrained matrix equation on signed integers--is also realized to show graphically the algorithm dynamics.

Bioinformatics, 1991

Genotype by environment interaction (GEI) is a relevant topic in many fields of applied biology. ... more Genotype by environment interaction (GEI) is a relevant topic in many fields of applied biology. Normal parametric approaches following analysis of variance are not suitable when large differences between error variances within environments are present. We illustrate an APL program following Yamada's approach to estimate GEIs and ancillary statistics when two environments are considered. The choice of APL environment is based on its value as a useful tool in algorithm implementation and problem-solving. Its use is suggested in biological applications where matrix algebra is involved and an understanding of the computing complexity of problems is generally not required.

Bioinformatics, 2000

Motivation: Molecular profiles (DNA fingerprints) may be used to allocate an individual of unknow... more Motivation: Molecular profiles (DNA fingerprints) may be used to allocate an individual of unknown membership to one among the known groups of a reference population. Time and costs of profile assessment may be reduced by identifying informative profile components (markers). Results: A genetic algorithm (GA) is proposed to identify promising candidate markers from a pilot experiment in which observations are supposed to be without measurement error. The analysis of simulated datasets suggests reasonable values for GA parameters and confirms that the GA finds components of the profile showing association with the considered groups. Our GA may be used to perform a first screening of candidate markers to be included in subsequent experiments. Availability: The 32-bit executable (Windows 95, 98 and NT) is available at http:// www.ds.unifi.it/ ∼stefanin/ bioinformatics.htm. Contact: stefanin@ds.unifi.it Supplementary Information: The algorithm is implemented for research purposes, i.e. a limited amount of input filtering and error messages are provided.

Artificial Intelligence in Medicine, 2012

Setting up clinical reports within hospital information systems makes it possible to record a var... more Setting up clinical reports within hospital information systems makes it possible to record a variety of clinical presentations. Directed acyclic graphs (Dags) offer a useful way of representing causal relations in clinical problem domains and are at the core of many probabilistic models described in the medical literature, like Bayesian networks. However, medical practitioners are not usually trained to elicit Dag features. Part of the difficulty lies in the application of the concept of direct causality before selecting all the causal variables of interest for a specific patient. We designed an automated interview to tutor medical doctors in the development of Dags to represent their understanding of clinical reports. Medical notions were analyzed to find patterns in medical reasoning that can be followed by algorithms supporting the elicitation of causal Dags. Clinical relevance was defined to help formulate only relevant questions by driving an expert's attention towards variables causally related to nodes already inserted in the graph. Key procedural features of the proposed interview are described by four algorithms. The automated interview comprises questions on medical notions, phrased in medical terms. The first elicitation session produces questions concerning the patient's chief complaints and the outcomes related to diseases serving as diagnostic hypotheses, their observable manifestations and risk factors. The second session focuses on questions that refine the initial causal paths by considering syndromes, dysfunctions, pathogenic anomalies, biases and effect modifiers. A case study concerning a gastro-enterological problem and one dealing with an infected patient illustrate the output produced by the algorithms, depending on the answers provided by the doctor. The proposed elicitation framework is characterized by strong consistency with medical background and by a progressive introduction of relevant medical topics. Revision and testing of the subjectively elicited Dag is performed by matching the collected answers with the evidence included in accepted sources of biomedical knowledge.

International Conference on Networking, Sensing and Control, 1999

Most of the approaches developed in the literature to elicit the a-priori distribution on Directe... more Most of the approaches developed in the literature to elicit the a-priori distribution on Directed Acyclic Graphs (DAGs) require a full specification of graphs. Nevertheless, expert's prior knowledge about conditional independence relations may be weak, making the elicitation task troublesome. Moreover, the detailed specification of prior distributions for structural learning is NP-Hard, (7), making the elicitation of large net- works

In Vitro Cellular & Developmental Biology - Animal

Bayesian Networks are increasingly used to represent conditional in- dependence relations among v... more Bayesian Networks are increasingly used to represent conditional in- dependence relations among variables and causal information in problem domains in which decisions are based on probabilistic reasoning. Structural learning is NP- hard therefore the database of observed cases must be often supplemented with search heuristics based on prior information. In this paper we present a software package for R, called

Alternatives to laboratory animals : ATLA, 2011

The human carcinogenicity evaluation of chemicals has a great impact on public health. In vitro m... more The human carcinogenicity evaluation of chemicals has a great impact on public health. In vitro methods, such as the cell transformation assay (CTA), allow for a fast and reliable assessment of the carcinogenic potential of a chemical compound in comparison with the standard two-year bioassay. The scoring and classification of foci in selected cell lines is performed, after staining, by light microscopy. Foci can be separated into three classes: type I, which are scored as non-transformed, and types II and III that are considered to include fully transformed foci. However, in a number of cases, even an expert is uncertain about the attribution of a focus to a given class, due to its mixed or intermediate nature. Here, we suggest a simple approach to classifying mixed or intermediate foci by exploiting the quantitative information available from images, which is captured by statistical descriptors. A quantitative index is proposed, to describe the degree of dissimilarity of mixed and...

Imaging, Manipulation, and Analysis of Biomolecules, Cells, and Tissues VIII, 2010

The Cell Transformation Assay (CTA) is one of the promising in vitro methods used to predict huma... more The Cell Transformation Assay (CTA) is one of the promising in vitro methods used to predict human carcinogenicity. The neoplastic phenotype is monitored in suitable cells by the formation of foci and observed by light microscopy after staining. Foci exhibit three types of morphological alterations: Type I, characterized by partially transformed cells, and Types II and III considered to have

Lecture Notes in Computer Science, 2007

Most of the approaches developed in the literature to elicit the a-priori distribution on Directe... more Most of the approaches developed in the literature to elicit the a-priori distribution on Directed Acyclic Graphs (DAGs) require a full specification of graphs. Nevertheless, expert’s prior knowledge about conditional independence relations may be weak, making the elicitation task troublesome. Moreover, the detailed specification of prior distributions for structural learning is NP-Hard, making the elicitation of large networks impractical. This

Biometrical Letters, 2013

Journal of the Royal Statistical Society: Series A (Statistics in Society), 2003

The evaluation of nuclear DNA evidence for identification purposes is performed here taking accou... more The evaluation of nuclear DNA evidence for identification purposes is performed here taking account of the uncertainty about population parameters. Graphical models are used to detail the hypotheses being debated in a trial with the aim of obtaining a directed acyclic graph. Graphs also clarify the set of evidence that contributes to population inferences and they also describe the conditional independence structure of DNA evidence. Numerical illustrations are provided by re-examining three case-studies taken from the literature. Our calculations of the weight of evidence differ from those given by the authors of case-studies in that they reveal more conservative values.

Journal of Microscopy, 2009

Carcinogenesis is a multi-step process involving genetic alterations and non-genotoxic mechanisms... more Carcinogenesis is a multi-step process involving genetic alterations and non-genotoxic mechanisms. The in vitro cell transformation assay allows the monitoring of the neoplastic phenotype by foci formation in suitable cells (e.g. C3H10T1/2 mouse embryo fibroblasts) showing aberrant morphology of massive build-up, polar and multi-layered densely stained cells. The classification of transformed foci in C3H cells relies on light microscopy scoring by a trained human expert based on standard rules. This procedure is time-consuming and prone, in some cases, to subjectivity, thereby leading to possible over- or under-estimation of the carcinogenic potential of tested compounds. Herewith we describe the in vitro neoplastic transformation induced by B[a]P and CdCl2, and the development of a foci classifier based on image analysis and statistical classification. The image analysis system, which relies on 'spectrum enhancement', is quantitative and extracts descriptors of foci texture and structure. The statistical classification method is based on the Random Forest algorithm. We obtained a classifier trained by using expert's supervision with a 20% classification error. The proposed method could serve as a basis to automate the in vitro cell transformation assay.

Journal of Applied Microbiology, 2007

To assess the effect of different concentrations of gossypol (0, 2, 4, 10 and 20 mg l(-1)) in com... more To assess the effect of different concentrations of gossypol (0, 2, 4, 10 and 20 mg l(-1)) in combination with NaCl and Na(2)SO(4) (20 mS cm(-1)) on the conidial germination and viability of Fusarium oxysporum f.sp. vasinfectum (Fov). A multinomial logistic model was developed to estimate the germination probability of Fov. The inhibitory effect was markedly evident at the two highest concentrations of gossypol; it varied among the isolates tested and with time, and it was attenuated by the presence of sodium salts. The inhibition was temporary as the germination probability increased after 8 h. Fluorescent staining revealed that gossypol either killed the conidia or retarded the elongation of the germ tubes. Fov showed the ability to overcome gossypol inhibition over time, and the inhibitory effect is reduced under saline conditions. Differential responses among Fov isolates to the presence of gossypol suggest that gossypol tolerance is genetically determined in the pathogen. This study suggests that selecting for high plant gossypol cultivars would have minimal effect on the overall Fov resistance of cotton. A new statistical model was developed to explore the statistical significance of plant-pathogen interactions.

Euphytica, 1995

Potato plantlets derived from in vitro propagation of three cultivars known for their field resis... more Potato plantlets derived from in vitro propagation of three cultivars known for their field resistance (Cruza 148 and BR-63.65) or susceptibility (Drsirre) to Pseudomonas solanacearum E.E Smith were artificially inoculated under controlled conditions. The aim of this work was to determine the optimal inoculum concentration and the best observation period in which the cultivars would show different reactions to bacterial infection as expected on the basis of their field performance.

Canadian Journal of Forest Research, 1995

To determine whether cypress bark canker, caused by the fungus Seiridium cardinale (Wag.) Sutton ... more To determine whether cypress bark canker, caused by the fungus Seiridium cardinale (Wag.) Sutton & Gibson, affects height and diameter growth of common cypress (Cupressus sempervirens L.), these parameters were measured over 4 years on inoculated and uninoculated ramets of 50 cypress clones at two locations in Italy. An analysis of covariance, using the pre-trial diameter and height of the ramets as the covariates, showed that the disease had no effect on either height or diameter growth.

Bioinformatics, 1993

Problem-solving and modelling within a biological context often need a level of descriptive accur... more Problem-solving and modelling within a biological context often need a level of descriptive accuracy that is unlikely to be capable of analytical treatment, especially if the mathematical background of the biologist is poor. Furthermore solver-model maintenance is often difficult without the availability of trained specialists. Better prospects are found in the genetic algorithm field. Genetic algorithms are a set of procedures formulated to solve complex problems without specifying rules for intermediate steps. This approach becomes feasible performing a Monte Carlo simulation of the natural evolution process, in which population improvement (search for solutions) in a considered environment (the specific problem domain) is achieved by following the genetic paradigm. Starting with a randomly constituted sample of individuals, drawn from the population of admissible values and expressed as binary strings, random mating brings about individuals of the next generation. Parents are chosen with a greater probability as the number of constraints violated by each individual becomes smaller. During the constitution of each generation the presence of some genetic operators causes the improvement of population diversity and its maintenance. Genetic operators are simple string transformation rules, generally independent of a specific context. We have developed the constant core of a minimal genetic algorithm, from which can be derived genetic problem-solvers in specific domains. An applicative example--a constrained matrix equation on signed integers--is also realized to show graphically the algorithm dynamics.

Bioinformatics, 1991

Genotype by environment interaction (GEI) is a relevant topic in many fields of applied biology. ... more Genotype by environment interaction (GEI) is a relevant topic in many fields of applied biology. Normal parametric approaches following analysis of variance are not suitable when large differences between error variances within environments are present. We illustrate an APL program following Yamada's approach to estimate GEIs and ancillary statistics when two environments are considered. The choice of APL environment is based on its value as a useful tool in algorithm implementation and problem-solving. Its use is suggested in biological applications where matrix algebra is involved and an understanding of the computing complexity of problems is generally not required.

Bioinformatics, 2000

Motivation: Molecular profiles (DNA fingerprints) may be used to allocate an individual of unknow... more Motivation: Molecular profiles (DNA fingerprints) may be used to allocate an individual of unknown membership to one among the known groups of a reference population. Time and costs of profile assessment may be reduced by identifying informative profile components (markers). Results: A genetic algorithm (GA) is proposed to identify promising candidate markers from a pilot experiment in which observations are supposed to be without measurement error. The analysis of simulated datasets suggests reasonable values for GA parameters and confirms that the GA finds components of the profile showing association with the considered groups. Our GA may be used to perform a first screening of candidate markers to be included in subsequent experiments. Availability: The 32-bit executable (Windows 95, 98 and NT) is available at http:// www.ds.unifi.it/ ∼stefanin/ bioinformatics.htm. Contact: stefanin@ds.unifi.it Supplementary Information: The algorithm is implemented for research purposes, i.e. a limited amount of input filtering and error messages are provided.

Artificial Intelligence in Medicine, 2012

Setting up clinical reports within hospital information systems makes it possible to record a var... more Setting up clinical reports within hospital information systems makes it possible to record a variety of clinical presentations. Directed acyclic graphs (Dags) offer a useful way of representing causal relations in clinical problem domains and are at the core of many probabilistic models described in the medical literature, like Bayesian networks. However, medical practitioners are not usually trained to elicit Dag features. Part of the difficulty lies in the application of the concept of direct causality before selecting all the causal variables of interest for a specific patient. We designed an automated interview to tutor medical doctors in the development of Dags to represent their understanding of clinical reports. Medical notions were analyzed to find patterns in medical reasoning that can be followed by algorithms supporting the elicitation of causal Dags. Clinical relevance was defined to help formulate only relevant questions by driving an expert's attention towards variables causally related to nodes already inserted in the graph. Key procedural features of the proposed interview are described by four algorithms. The automated interview comprises questions on medical notions, phrased in medical terms. The first elicitation session produces questions concerning the patient's chief complaints and the outcomes related to diseases serving as diagnostic hypotheses, their observable manifestations and risk factors. The second session focuses on questions that refine the initial causal paths by considering syndromes, dysfunctions, pathogenic anomalies, biases and effect modifiers. A case study concerning a gastro-enterological problem and one dealing with an infected patient illustrate the output produced by the algorithms, depending on the answers provided by the doctor. The proposed elicitation framework is characterized by strong consistency with medical background and by a progressive introduction of relevant medical topics. Revision and testing of the subjectively elicited Dag is performed by matching the collected answers with the evidence included in accepted sources of biomedical knowledge.