Use of neural networks to model complex immunogenetic associations of disease: human leukocyte antigen impact on the progression of human immunodeficiency virus infection.

Complex immunogenetic associations of disease involving a large number of gene products are difficult to evaluate with traditional statistical methods and may require complex modeling. The authors evaluated the performance of feed-forward backpropagation neural networks in predicting rapid progression to acquired immunodeficiency syndrome (AIDS) for patients with human immunodeficiency virus (HIV) infection on the basis of major histocompatibility complex variables. Networks were trained on data from patients from the Multicenter AIDS Cohort Study (n = 139) and then validated on patients from the DC Gay cohort (n = 102). The outcome of interest was rapid disease progression, defined as progression to AIDS in <6 years from seroconversion. Human leukocyte antigen (HLA) variables were selected as network inputs with multivariate regression and a previously described algorithm selecting markers with extreme point estimates for progression risk. Network performance was compared with that of logistic regression. Networks with 15 HLA inputs and a single hidden layer of five nodes achieved a sensitivity of 87.5% and specificity of 95.6% in the training set, vs. 77.0% and 76.9%, respectively, achieved by logistic regression. When validated on the DC Gay cohort, networks averaged a sensitivity of 59.1% and specificity of 74.3%, vs. 53.1% and 61.4%, respectively, for logistic regression. Neural networks offer further support to the notion that HIV disease progression may be dependent on complex interactions between different class I and class II alleles and transporters associated with antigen processing variants. The effect in the current models is of moderate magnitude, and more data as well as other host and pathogen variables may need to be considered to improve the performance of the models. Artificial intelligence methods may complement linear statistical methods for evaluating immunogenetic associations of disease.

[1]  Anders Krogh,et al.  Introduction to the theory of neural computation , 1994, The advanced book program.

[2]  N Morling,et al.  HLA and disease. , 1977, Transplantation proceedings.

[3]  Patrick van der Smagt,et al.  Introduction to neural networks , 1995, The Lancet.

[4]  Etienne Barnard,et al.  Avoiding false local minima by proper initialization of connections , 1992, IEEE Trans. Neural Networks.

[5]  J. Goedert,et al.  Serum HIV-1 RNA levels and time to development of AIDS in the Multicenter Hemophilia Cohort Study. , 1996, JAMA.

[6]  J. Phair,et al.  The Multicenter AIDS Cohort Study: rationale, organization, and selected characteristics of the participants. , 1987, American journal of epidemiology.

[7]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[8]  Robert A. Jacobs,et al.  Increased rates of convergence through learning rate adaptation , 1987, Neural Networks.

[9]  A Hart,et al.  Evaluating black-boxes as medical decision aids: issues arising from a study of neural networks. , 1990, Medical informatics = Medecine et informatique.

[10]  P. Lapuerta,et al.  Use of neural networks in predicting the risk of coronary artery disease. , 1995, Computers and biomedical research, an international journal.

[11]  David W. Hosmer,et al.  Applied Logistic Regression , 1991 .

[12]  J. Trowsdale,et al.  The TAP complex influences allorecognition of class II MHC molecules. , 1996, Human immunology.

[13]  Kevin N. Gurney,et al.  An introduction to neural networks , 2018 .

[14]  J. Goedert,et al.  Concordance of human leukocyte antigen haplotype‐sharing, CD4 decline and AIDS in hemophilic siblings , 1995, AIDS.

[15]  P. Terasaki,et al.  HLA and Disease Associations , 1985, Springer New York.

[16]  W. Baxt.  Application of artificial neural networks to clinical medicine , 1995, The Lancet.

[17]  J J Goedert,et al.  Genetic Restriction of HIV-1 Infection and Progression to AIDS by a Deletion Allele of the CKR5 Structural Gene , 1996, Science.

[18]  Geoffrey E. Hinton,et al.  Learning representations by back-propagating errors , 1986, Nature.

[19]  J. Goedert,et al.  Decreased helper T lymphocytes in homosexual men. I. Sexual contact in high-incidence areas for the acquired immunodeficiency syndrome. , 1985, American journal of epidemiology.

[20]  P. Cresswell,et al.  Presentation of viral antigen by MHC class I molecules is dependent on a putative peptide transporter heterodimer , 1992, Nature.

[21]  H. Maeta,et al.  Prediction of the early prognosis of the hepatectomized patient with hepatocellular carcinoma with a neural network. , 1995, Computers in biology and medicine.

[22]  W. Baxt.  Use of an artificial neural network for the diagnosis of myocardial infarction. , 1991, Annals of internal medicine.

[23]  P. Parham,et al.  The Bw4 public epitope of HLA-B molecules confers reactivity with natural killer cell clones that express NKB1, a putative HLA receptor , 1995, The Journal of experimental medicine.

[24]  R. Dybowski,et al.  Prediction of outcome in critically ill patients using artificial neural network synthesised by genetic algorithm , 1996, The Lancet.

[25]  R. Kaslow,et al.  The role of the major histocompatibility complex in human immunodeficiency virus infection--ever more complex? , 1994, Journal of Infectious Diseases.

[26]  T. Chalmers,et al.  Human leukocyte antigen associations of epidemic Kaposi's sarcoma. , 1995, AIDS.

[27]  J. Goedert,et al.  Influence of combinations of human major histocompatibility complex genes on the course of HIV–1 infection , 1996, Nature Medicine.

[28]  J V Tu,et al.  Advantages and disadvantages of using artificial neural networks versus logistic regression for predicting medical outcomes. , 1996, Journal of clinical epidemiology.