The determination of gene expression is a very common scientific method used in modern laboratory for a variety of applications. One of the most popular is the real time PCR, a quantitative modification of the classic PCR method where the increase of the amplify nucleic acid is examined cycle by cycle after every amplification step. The analysis of the PCR productduring the amplification process allows to compare the initial amount of cDNA synthesized from isolated RNA and calculate the number of particular RNA copies present in examined material. In spite of obvious advantages of real time PCR there are also some inconveniences of this method. First of all, there is no possibility of analyzing more than one gene in a single reaction mixture. It is limited by the necessity of design and usage of different pairs of primers for each analyzed gene. Therefore, it is necessary to predict the cell, tissue or organism response for applied treatment, examined condition, etc. The development of microarray methods enables to overcome these problems and parallel analyze all known genes in the single sample at the same time. There is no need to predict which gene expression might be changed under studied conditions because the microarray data is a comprehensive pattern of the expression of all known genes, which probes are implemented on the microchip surface. Although the microarray data is an excellent method for gene expression comparison, the estimation of the extent of change fold is not very precise and usually is confirmed and determined by real time PCR with respect to selected genes. The method which combines the quantitative precision of real time PCR and the possibility to analyze broad spectrum of the genes is a deep sequencing method also called next generation sequencing. It is a new method developed for the analysis of the whole RNA isolated from a sample without the need to design primers and thus any knowledge of expressed genes sequence. The advantages of this method include the possibility of finding unexpected expression of completely unknown DNA fragments, alternative splicing variants of the genes and differences in DNA sequence. The deep sequencing provides an extremely large amount of information, much more than microarray data, and to analyse it new bioinformatics methods and tools especially designed for this purpose are required.
The methods used in biomedical research are becoming inadequate to meet current challenges. Frequently occurring problem is the need to find the differentiation tests according to phenotypic features or the particular phenomenon. Previously used morphological evaluation or other laboratory tests many times do not allow for adequate determination of differentiating attributes. In recent years there has been considerable scientific and technological progress in the fields such as genomics, transcriptomics, proteomics and metabolomics, which allow to move the search area into the molecular level. It allows the use of advanced molecular techniques such as PCR or oligonucleotide microarrays and thus allows to compare the gene expression profiles of different types of cells and tissues. The microarray experiment data allow to determine the correlation between the expression of selected genes or even entire genotypes of the phenotypic features, characterizing the studied group. The collected data can not be analyzed using traditional statistical methods, since the number of cases is much higher than the number of considered attributes. For this reason, new statistical methods and procedures are used for microarray data analysis which may focus on theoretical or practical aspects. Theoretical aspect is related to the selection of specific genes expression, finding the ontology or metabolic pathways that are associated with the analyzed phenomenon. The practical aspect can be the creation of a predictive model that can allow to predict the specific phenonon occurrence in the future during the studies of new patients. Microarray experiments and analysis of the obtained results begin new chapters of particular phenomena investigation, which is another big boost in the biomedical sciences development.
Infertility is a serious social problem. Very often the only treatment possibility are IVF methods. This study explores the possibility of outcome prediction in the early stages of treatment. The data, collected from the previous treatment cycles, were divided into four subsets, which corresponded to the selected stages of treatment. On each such subset, sophisticated data mining analysis was carried out, with appropriate imputations and classification procedures. The obtained results indicate that there is a possibility of predicting the final outcome at the beginning of treatment.
The effectiveness of IVF ICSI/ET infertility treatment depends on many factors. Their identification and classification of individual cases remains a difficult task. This paper presents application of feature selection algorithm MSIMBAF2 and associated kNN classifier to analyze the data set containing results of the infertility treatment process.
The IVF ET method is a scientifically recognized infertility treat- ment method. The problem, however, is this method’s unsatisfactory efficiency. This calls for a more thorough analysis of the information available in the treat- ment process, in order to detect the factors that have an effect on the results, as well as to effectively predict result of treatment. Classical statistical methods have proven to be inadequate in this issue. Only the use of modern methods of data mining gives hope for a more effective analysis of the collected data. This work provides an overview of the new methods used for the analysis of data on infertility treatment, and formulates a proposal for further directions for research into increasing the efficiency of the predicted result of the treatment process.
One of the most effective methods of infertility treatment is in vitro fertilization (IVF). Effectiveness of the treatment, as well as classification of the data obtained from it, is still an ongoing issue. Classifiers obtained so far are powerful, but even the best ones do not exhibit equal quality concerning possible treatment outcome predictions. Usually, lack of pregnancy is predicted far too often. This creates a constant need for further exploration of this issue. Careful use of different classification methods can, however, help to achieve that goal.
JavaScript is turned off in your web browser. Turn it on to take full advantage of this site, then refresh the page.