Identifying genetic variation that underlies complex phenotypes is an important problem in genetics. Despite great progress in the molecular genetic methods to identify both rare and common variants in the whole genome, there has been considerably less progress in the refinement of phenotypes. Many biological traits are heterogeneous, which substantially limits the identification of genotype-phenotype correlations or genetic linkage results. Differentiating homogeneous subtypes of a complex phenotype can improve the identification of genetic factors for many complex biological systems.
There has been a great need for statistically rigorous methods to derive heterogeneous subphenotypes. Currently, the most sophisticated subtyping methods perform unsupervised cluster analysis on phenotypic features, and the most sophisticated subtype analysis methods (once subtypes are defined) perform multinomial analysis over multiple subtypes. This special issue aims to provide a forum for researchers who work on the development of novel computational methods for quantitative subtyping of complex traits and related methods to improve genetic analysis (association, sequencing or linkage analysis) for subtype analysis, which will translate into personalised diagnosis and therapy.
There are two directions of subtyping: molecular subtyping and clinical syndrome-level subtyping. Molecular subtyping differentiates subjects with a disease by gene expression signatures or other molecular-level factors. Clinical subtyping is used to identify homogeneous traits at the clinical syndrome level. For many complex diseases, such as mental illness, subtyping based on clinical features might be the most feasible study. For cancer, the former approach which mainly uses clinical syndrome information, however, can be ineffective for discriminating multiple illnesses with overlapping syndromes. The diagnosis or subtyping of complex diseases depends on multiple factors including genomic, epigenomic, clinical and environmental changes, necessitating multiscale approaches.
Recently, the progress of genomic technology has opened new avenues. The multi-omics data from next generation sequencing (NGS), epigenomics, transcriptomics, functional genomics, proteomics, and molecular and cell biology provide multiple and complementary information, which can be used as novel biomarkers for improved subtyping. However, the mining and integrating these biomarkers for subtyping of complex diseases and complex phenotypes pose significant challenges. This issue will publish papers selected from the subtyping workshop jointly held with IEEE International Conference on Bioinformatics and Biomedicine 2013, but welcomes contributions from any other researcher discussing the challenges outlined above and related challenges and solutions.
Suitable topics include but not limited to:
- Association analysis, QTL, eQTL with multiple phenotypes or traits jointly
- QTL or eQTL with multiple quantitative traits that are subtypes of a complex phenotype
- Studies that prove the advantages of subtyping and subtype analysis
- Case studies of various complex diseases related to the analysis of multiple disease phenotypes
- Quantitative search of subtypes from electronic medical records related to a complex disease
- Cancer subtype discovery based on tumor gene expression, somatic mutations, or epigenome data
- The integration of multi-omics data for subtype analysis
- The identification of novel biomarkers from NGS platforms that can be used to improve subtyping
- Novel mathematical and computational models such as sparse regression and cluster analysis for the classification and identification of phenotypic subtypes
Due date for paper submission: 20 February, 2014