【摘要】 本文所涉及的课题是白细胞分型研究中的应用软件之一,文中对白细胞分型研究中血清特异性鉴定的数学原理作了阐述,对反应结果的尾巴部份进行聚类,以确定可能存在的新的特异性,并对血清图绘制的原理作了介绍。更多还原

【Abstract】 This paper presents the mathematical principles of the computer programs for identifying serum specificities and drawing serographs in the study of HIA system. Specificities of unknown cytotoxic sera were assigned with a well-defined lymphocyte panel serving as the reference. Unknown cytotoxic sera were first tested with the lymphocyte panel. A series of four-cell tables could be constructed when comparing the reaction pattern of one serum with the distribution pattern of antigens included in the lymphocyte panel. From these four-cell tables correlation coefficients and S(score) values could be calculated. Highly significant value of correlation coefficient indicates the presence of specific HLA antibody in the serum. S value is defined as the logarithm of probability ratio (H) between the probability of having the specificity and that of not having the specificity when the reaction pattern of a serum is compared with the antigen distribution pattern in the lymphocyte panel. The greater the deviation from zero of S value is, the greater the probability of having the corresponding antibody in the serum will be. When a serum has tail reactions in addition to the principal specificity, the reactions for the principal specificity could be erased by the computer and the remaining reaction pattern could be treated as a new serum. In this way a multi-specific serum can be identified. After the identification of specificities for a group of sera, those with identical or related specificities could be clustered. A clear serograph forthe cluster could then be prepared by the computer by way of rearranging the order of the sera and of the cells. The paper presents the procedures with which the rearrangements can be done.更多还原