Molecular signatures of cancer, e.g., genes or microRNAs (miRNAs), have been recognized very important in predicting the occurrence of cancer. From gene-expression and miRNA-expression data, the challenge of identifying molecular signatures lies in the huge number of molecules compared to the small number of samples. To address this issue, in this paper, we propose a heuristic algorithm to identify molecular signatures, termed HAMS, for cancer diagnosis by modeling it as a multi-objective optimization problem. In the proposed HAMS, an elitist-guided individual update strategy is proposed to obtain a small number of molecular signatures, which are closely related with cancer and contain less redundant signatures. Experimental results demonstrate that the proposed HAMS achieves superior performance over seven state-of-the-art algorithms on both gene-expression and miRNA-expression datasets. We also validate the biological significance of the molecular signatures obtained by the proposed HAMS through biological analysis.

中文翻译：

---

用于识别癌症分子特征的启发式算法。

已经认识到癌症的分子标记，例如基因或微小RNA（miRNA），在预测癌症的发生中非常重要。从基因表达和miRNA表达数据来看，鉴定分子标记的挑战在于，与少量样品相比，存在大量的分子。为了解决这个问题，在本文中，我们提出了一种启发式算法，通过将其建模为多目标优化问题来识别被称为HAMS的分子特征，以进行癌症诊断。在提出的HAMS中，提出了一种由专家指导的个体更新策略，以获取少量与癌症密切相关且包含较少冗余签名的分子签名。实验结果表明，所提出的HAMS在基因表达和miRNA表达数据集上均优于7种最新算法。我们还通过生物学分析验证了所提出的HAMS获得的分子标记的生物学意义。