Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Correspondence
  • Published:

Machine learning applied to gene expression analysis of T-lymphocytes in patients with cGVHD

Bone Marrow Transplantation volume 55, pages 1668–1670 (2020)Cite this article

Subjects

Chronic graft-vs-host disease (cGVHD) remains the major cause of late morbility and mortality in non-relapsing patients after allogeneic hematopoietic stem cell transplantation (allo-HSCT) [1]. Key role in the development of cGVHD is mainly displayed by donor pro-inflammatory T-cell subpopulations [2]. Beyond standard treatments, cell-based approaches and novel agents as ibrutinib or ruxolitinib have recently emerged in second-line setting [3]. In attempt to distinguish between normal and pathological immune processes, promising biofluids biomarkers instead of gene expression profiles (GEP) are being described, although they have not been validated yet [4]. Given the massive data generated in these assays, machine learning (ML)-inspired approaches can allow the processing of multidimensional data and have shown to uncover clinical patterns in several human diseases [5]. In this study, we applied a set of ML algorithms on GEP data, in order to generating a cGVHD diagnostic model based on the description of gene classifiers.

A total of 19 patients who had underwent HLA-matched allo-HSCT and 10 healthy donors (Helathy Control group) were included (Supplementary Table 1). At the time of sampling and according to NIH consensus 2014 [6], nine patients (cGVHD group) showed chronic GVHD symptoms (four mild, four severe and one moderate) whereas ten patients did not show any cGVHD signs (No-cGVHD group). RNAs obtained from circulating T cells were used to perform microarray assay. Then, we proceeded with selection of relevant RNAs by using ML strategies considering the limitation of the sample size as a starting point. Hence, we designed a procedure consisting on a two-staged feature selection, followed by a Random Forest (RF) [7] model fitting. Based on pioneering works where successfully applied decision trees or RF on gene expression analysis for tumors categorization, by using a small set of selected genes [8], we decided to employ RF in this study. (Supplementary Methodology). Unbiasedly, we proceed to perform our study blindly and by unsupervised methods. Thus, in the preprocessing of our massive data obtained from microarrays, we applied Expectation–Maximization algorithm, an unsupervised method with high capacity of managing missing information dataset [9] for discarding irrelevant genes. In a second step, Boruta algorithm [10] helps us to uncover a small set of relevant mRNAs that categorized our patients (Supplementary Fig. 1).

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Gene clustering and their influence in each group.

References

  1. MacDonald KP, Hill GR, Blazar BR. Chronic graft-versus-host disease: biological insights from preclinical and clinical studies. Blood. 2017;129:13–21.

    Article  CAS  Google Scholar 

  2. Finke J, Bethge WA, Schmoor C, Ottinger HD, Stelljes M, Zander AR, et al. Standard graft-versus-host disease prophylaxis with or without anti-T-cell globulin in haematopoietic cell transplantation from matched unrelated donors: a randomised, open-label, multicentre phase 3 trial. Lancet Oncol. 2009;10:855–64.

    Article  CAS  Google Scholar 

  3. Im A, Hakim FT, Pavletic SZ. Novel targets in the treatment of chronic graft-versus-host disease. Leukemia. 2017;31:543–54.

    Article  CAS  Google Scholar 

  4. Yu J, Storer BE, Kushekhar K, Abu Zaid M, Zhang Q, Gafken PR, et al. Biomarker panel for chronic graft-versus-host disease. J Clin Oncol. 2016;34:2583–90.

    Article  CAS  Google Scholar 

  5. Gandelman JS, Byrne MT, Mistry AM, Polikowsky HG, Diggins KE, Chen H, et al. Machine learning reveals chronic graft-versus-host disease phenotypes and stratifies survival after stem cell transplant for hematologic malignancies. Haematologica. 2019;104:189–96.

    Article  CAS  Google Scholar 

  6. Lee SJ. Classification systems for chronic graft-versus-host disease. Blood. 2017;129:30–37.

    Article  CAS  Google Scholar 

  7. Breiman L. Randon forest. Mach Learn. 2001;45:5–32.

    Google Scholar 

  8. Diaz-Uriarte R, de Andres S. Gene selection and classification of microarray data using random forest. BMC Bioinforma. 2006;7:1–13.

    Article  Google Scholar 

  9. Xiaodan F, Yuan Y, Jun SL. The EM algorithm and the rise of computational biology. Stat Sci. 2010;25:476–91.

    Article  Google Scholar 

  10. Kursa MB, Rudnicki WR. Feature selection with Boruta Package. J Stat Soft. 2010;36:2–13.

    Article  Google Scholar 

  11. Jiang P, Gu S, Pan D, Fu J, Sahu A, Hu X, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy reponse. Nat Med. 2018;4:1550–58.

    Article  Google Scholar 

  12. Horn S, Leonardelli S, Sucker A. Tumor CDKN2A-associated JAK2 loss and susceptibility to immunotherapy resistance. JNCI J Natl Cancer Inst. 2018;110:677–81.

    Article  CAS  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Instituto Maimonides Investigación Biomédica, IMIBIC/University of Córdoba, Córdoba, Spain

    Juana Serrano-López, José Luis Fernández, Josefina Serrano, Carmen Martínez-Losada, Carmen Martín & Joaquín Sánchez-García

  2. Experimental Hematology Lab, IIS-Fundación Jiménez Díaz, UAM, Madrid, Spain

    Juana Serrano-López

  3. Centro de Investigación del Cancer, Universidad de Salamanca, CSIC, Salamanca, Spain

    Eva Lumbreras & Jesús M. Hernández-Rivas

  4. Hematology Department, Hospital Universitario Reina Sofía, Córdoba, Spain

    Josefina Serrano, Carmen Martínez-Losada, Carmen Martín & Joaquín Sánchez-García

  5. Hospital Universitario de Salamanca IBSAL, Salamanca, Spain

    Jesús M. Hernández-Rivas

Authors
  1. Juana Serrano-López
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. José Luis Fernández
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Eva Lumbreras
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Josefina Serrano
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Carmen Martínez-Losada
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Carmen Martín
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Jesús M. Hernández-Rivas
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Joaquín Sánchez-García
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Juana Serrano-López.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Serrano-López, J., Fernández, J.L., Lumbreras, E. et al. Machine learning applied to gene expression analysis of T-lymphocytes in patients with cGVHD. Bone Marrow Transplant 55, 1668–1670 (2020). https://doi.org/10.1038/s41409-020-0848-y

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41409-020-0848-y

Search

Advanced search

Quick links