陆军军医大学(第三军医大学)军事预防系医学系军队卫生统计学教研室,重庆
WU Yazhou, E-mail: asiawu@tmmu.edu.cn
Received:16 December 2024,
Revised:29 January 2025,
Published:30 March 2025
移动端阅览
李佳芮,钱力,沈俊杰,等.基于转录组学和变分自编码器的癌症分期诊断研究[J].陆军军医大学学报,2025,47(06):613-622.DOI:10.16016/j.2097-0927.202412079.
LI J R,QIAN L,SHEN J J,et al.Cancer staging diagnosis based on transcriptomics and variational autoencoder[J].J Army Med Univ,2025,47(06):613-622.DOI:10.16016/j.2097-0927.202412079.
目的
2
旨在对 10 种癌症转录组学数据开展深度分析与特征提取,进而实现对癌症样本的分期诊断。
方法
2
在UCSC Xena网站收集发病率最高的10种癌症转录组学数据(包含4 938个样本和59 428个基因),以变分自编码器为基础,利用特征重要性排序思想,通过引入掩码算法和增量特征选择方法(incremental feature selection,IFS),构建了增量特征排序选优变分自编码器(incremental feature ranking and selection variational autoencoder,IFRSVAE);与随机森林(random forest,RF)、支持向量机(support vector machine
SVM)和极限梯度提升树(eXtreme gradient boosting,XGboost)结合测试了该方法的性能,并同其他方法进行比较。
结果
2
提取了21个特征用于后续分类,较于传统变分自编码器、递归特征消除和Lasso回归模型,IFRSVAE模型在3种分类器上均取得了较好性能(AUC值最高,其余指标也表现良好);其中IFRSVAE-RF表现最好,AUC达到了85.49%(95%
CI
:83.24%~87.74%)。此外,还应用了沙普利加性解释(Shapley additive explanations, SHAP)模型对特征贡献进行展示。
结论
2
本研究探索并验证了IFRSVAE在特征提取方面具备一定成效。基于此构建的IFRSVAE-RF模型,在癌症分期诊断任务中展现出较好的表现,为深度学习在癌症分期诊断方法的研究方向上,提供了一种可供参考的新思路。
Objective To conduct an in-depth analysis and feature extraction of the transcriptomics data of 10 types of cancers in order to realize the staging diagn
osis of cancer samples. Methods The transcriptomics data of the top 10 cancers having the highest incidence were amassed from the UCSC Xena website
which comprised 4 938 samples and 59 428 genes. With the aid of variational autoencoder
we developed an incremental feature ranking and selection variational autoencoder (IFRSVAE) based on feature importance ranking and incorporating the masking algorithm and the Incremental Feature Selection (IFS). Subsequently
the performance efficiency of our IFRSVAE model was evaluated in conjunction with Random Forest (RF)
Support Vector Machine(SVM)
and eXtreme Gradient Boosting (XGboost)
and it was also compared with other methods. Results Our research extracted 21 features for the ensuing classification. In comparison to the conventional variational autoencoder
recursive feature elimination
and Lasso regression models
the IFRSVAE model attained more favorable performance across all 3 classifiers (highest AUC value
and well performed other indicators). Notably
the IFRSVAE-RF exhibited the most outstanding performance
with an AUC value reaching 85.49% (95%
CI
: 83.24%~87.74%). Moreover
Shapley additive explanations(SHAP) interpretable model illustrated well contributions of the features in our model. Conclusion Our developed IFRSVAE shows certain effectiveness in feature extraction. The constructed IFRSVAE-RF model demonstrates relatively good performance in the task of cancer staging diagnosis
which providing a new and referable idea for research orientation of deep-learning-based diagnostic methods for cancer staging.
QI J L , LI M L , WANG L J , et al . National and subnational trends in cancer burden in China, 2005-20: an analysis of national mortality surveillance data [J]. Lancet Public Health , 2023 , 8 ( 12 ): e943 - e955 . DOI: 10.1016/S2468-2667(23)00211-6 http://dx.doi.org/10.1016/S2468-2667(23)00211-6 .
BRAY F , LAVERSANNE M , WEIDERPASS E , et al . The ever-increasing importance of cancer as a leading cause of premature death worldwide [J]. Cancer , 2021 , 127 ( 16 ): 3029 - 3030 . DOI: 10.1002/cncr.33587 http://dx.doi.org/10.1002/cncr.33587 .
MA B S , MENG F Y , YAN G , et al . Diagnostic classification of cancers using extreme gradient boosting algorithm and multi-omics data [J]. Comput Biol Med , 2020 , 121 : 103761 . DOI: 10.1016/j.compbiomed.2020.103761 http://dx.doi.org/10.1016/j.compbiomed.2020.103761 .
KAUR R , BHARDWAJ A , GUPTA S . Cancer treatment therapies: traditional to modern approaches to combat cancers [J]. Mol Biol Rep , 2023 , 50 ( 11 ): 9663 - 9676 . DOI: 10.1007/s11033-023-08809-3 http://dx.doi.org/10.1007/s11033-023-08809-3 .
AMIN M B , GREENE F L , EDGE S B , et al . The eighth edition AJCC cancer staging manual: continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging [J]. CA Cancer J Clin , 2017 , 67 ( 2 ): 93 - 99 . DOI: 10.3322/caac.21388 http://dx.doi.org/10.3322/caac.21388 .
DONG Y Y , YANG W K , WANG J W , et al . MLW-gcForest: a multi-weighted gcForest model towards the staging of lung adenocarcinoma based on multi-modal genetic data [J]. BMC Bioinformatics , 2019 , 20 ( 1 ): 578 . DOI: 10.1186/s12859-019-3172-z http://dx.doi.org/10.1186/s12859-019-3172-z .
CAI Z H , XU D , ZHANG Q , et al . Classification of lung cancer using ensemble-based feature selection and machine learning methods [J]. Mol Biosyst , 2015 , 11 ( 3 ): 791 - 800 . DOI: 10.1039/c4mb00659c http://dx.doi.org/10.1039/c4mb00659c .
ATHIRA K , GOPAKUMAR G . Breast cancer stage prediction: a computational approach guided by transcrip-tome analysis [J]. Mol Genet Genomics , 2022 , 297 ( 6 ): 1467 - 1479 . DOI: 10.1007/s00438-022-01932-z http://dx.doi.org/10.1007/s00438-022-01932-z .
ALBARADEI S , NAPOLITANO F , THAFAR M A , et al . Meta Cancer: a deep learning-based pan-cancer metastasis prediction model developed using multi-omics data [J]. Comput Struct Biotechnol J , 2021 , 19 : 4404 - 4411 . DOI: 10.1016/j.csbj.2021.08.006 http://dx.doi.org/10.1016/j.csbj.2021.08.006 .
MARTINEZ-GARCIA M , OLMOS P M . Handling ill-conditioned omics data with deep probabilistic models [J]. IEEE J Biomed Health Inform , 2023 , 27 ( 9 ): 4601 - 4610 . DOI: 10.1109/JBHI.2023.3279493 http://dx.doi.org/10.1109/JBHI.2023.3279493 .
KINGMA D P , WELLING M . Auto-encoding variational Bayes [EB/OL]. 2013: 1312 . 6114 . https://arxiv.org/abs/1312.6114v11 https://arxiv.org/abs/1312.6114v11 .
SUNG H , FERLAY J , SIEGEL R L , et al . Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries [J]. CA Cancer J Clin , 2021 , 71 ( 3 ): 209 - 249 . DOI: 10.3322/caac.21660 http://dx.doi.org/10.3322/caac.21660 .
MA B S , CHAI B J , DONG H , et al . Diagnostic classification of cancers using DNA methylation of paracancerous tissues [J]. Sci Rep , 2022 , 12 ( 1 ): 10646 . DOI: 10.1038/s41598-022-14786-7 http://dx.doi.org/10.1038/s41598-022-14786-7 .
BARBIERI M , LEE P K , BRIZI L , et al . Circumventing the curse of dimensionality in magnetic resonance fingerprinting through a deep learning approach [J]. NMR Biomed , 2022 , 35 ( 4 ): e4670 . DOI: 10.1002/nbm.4670 http://dx.doi.org/10.1002/nbm.4670 .
LIU S Y , WANG Z T , ZHU R H , et al . Three differential expression analysis methods for RNA sequencing: limma, EdgeR, DESeq2 [J]. J Vis Exp , 2021 (175): (175). DOI: 10.3791/62528 http://dx.doi.org/10.3791/62528 .
REMESEIRO B , BOLON-CANEDO V . A review of feature selection methods in medical applications [J]. Comput Biol Med , 2019 , 112 : 103375 . DOI: 10.1016/j.compbiomed. 2019.103375 http://dx.doi.org/10.1016/j.compbiomed.2019.103375 .
SHAHIDI L K , COLLINS L M , MAINSAH B O . Parameter tuning of time-frequency masking algorithms for reverberant artifact removal within the cochlear implant stimulus [J]. Cochlear Implants Int , 2022 , 23 ( 6 ): 309 - 316 . DOI: 10.1080/14670100.2022.2096182 http://dx.doi.org/10.1080/14670100.2022.2096182 .
CAO S C , LI J J , NELSON K P , et al . Coupled VAE: improved accuracy and robustness of a variational autoencoder [J]. Entropy (Basel) , 2022 , 24 ( 3 ): 423 . DOI: 10.3390/e24030423 http://dx.doi.org/10.3390/e24030423 .
ZHANG Y H , HUANG T , CHEN L , et al . Identifying and analyzing different cancer subtypes using RNA-seq data of blood platelets [J]. Oncotarget , 2017 , 8 ( 50 ): 87494 - 87511 . DOI: 10.18632/oncotarget.20903 http://dx.doi.org/10.18632/oncotarget.20903 .
HU J C , SZYMCZAK S . A review on longitudinal data analysis with random forest [J]. Brief Bioinform , 2023 , 24 ( 2 ): bbad002 . DOI: 10.1093/bib/bbad002 http://dx.doi.org/10.1093/bib/bbad002 .
HUANG S J , CAI N G , PACHECO P P , et al . Applications of support vector machine (SVM) learning in cancer genomics [J]. Cancer Genomics Proteomics , 2018 , 15 ( 1 ): 41 - 51 . DOI: 10.21873/cgp.20063 http://dx.doi.org/10.21873/cgp.20063 .
CHEN S J , ZHOU W J , TU J H , et al . A novel XGBoost method to infer the primary lesion of 20 solid tumor types from gene expression data [J]. Front Genet , 2021 , 12 : 632761 . DOI: 10.3389/fgene.2021.632761 http://dx.doi.org/10.3389/fgene.2021.632761 .
BENJAMIN K J M , KATIPALLI T , PAQUOLA A C M . dRFEtools: dynamic recursive feature elimination for omics [J]. Bioinformatics , 2023 , 39 ( 8 ): btad513 . DOI: 10.1093/bioinformatics/btad513 http://dx.doi.org/10.1093/bioinformatics/btad513 .
DAI P P , CHANG W F , XIN Z R , et al . Retrospective study on the influencing factors and prediction of hospitalization expenses for chronic renal failure in China based on random forest and LASSO regression [J]. Front Public Health , 2021 , 9 : 678276 . DOI: 10.3389/fpubh.2021.678276 http://dx.doi.org/10.3389/fpubh.2021.678276 .
ALAWAD M , ALJOUIE A , ALAMRI S , et al . Machine learning and deep learning techniques for optic disc and cup segmentation—A review [J]. Clin Ophthalmol , 2022 , 16 : 747 - 764 . DOI: 10.2147/OPTH.S348479 http://dx.doi.org/10.2147/OPTH.S348479 .
ZANG Z L , CHENG S H , XIA H C , et al . DMT-EV: an explainable deep network for dimension reduction [J]. IEEE Trans Vis Comput Graph , 2024 , 30 ( 3 ): 1710 - 1727 . DOI: 10.1109/TVCG.2022.3223399 http://dx.doi.org/10.1109/TVCG.2022.3223399 .
ZHENG Q W , NI Q Z , ZHU B , et al . PPDPF promotes lung adenocarcinoma progression via inhibiting apoptosis and NK cell-mediated cytotoxicity through STAT3 [J]. Oncogene , 2022 , 41 ( 36 ): 4244 - 4256 . DOI: 10.1038/s41388-022-02418-3 http://dx.doi.org/10.1038/s41388-022-02418-3 .
LIN B H , HUANG G , YUAN Z S , et al . RPS15 coordinates with CtIP to facilitate homologous recombination and confer therapeutic resistance in breast cancer [J]. Radiat Res , 2024 , 202 ( 5 ): 775 - 784 . DOI: 10.1667/RADE-24-00134.1 http://dx.doi.org/10.1667/RADE-24-00134.1 .
SHI J L , FU L , ANG Q , et al . Overexpression of ATP1B1 predicts an adverse prognosis in cytogenetically normal acute myeloid leukemia [J]. Oncotarget , 2016 , 7 ( 3 ): 2585 - 2595 . DOI: 10.18632/oncotarget.6226 http://dx.doi.org/10.18632/oncotarget.6226 .
ECKERT M A , COSCIA F , CHRYPLEWICZ A , et al . Proteomics reveals NNMT as a master metabolic regulator of cancer-associated fibroblasts [J]. Nature , 2019 , 569 ( 7758 ): 723 - 728 . DOI: 10.1038/s41586-019-1173-8 http://dx.doi.org/10.1038/s41586-019-1173-8 .
HONG W C , LEE D E , KANG H W , et al . CD74 promotes a pro-inflammatory tumor microenvironment by inducing S100A8 and S100A9 secretion in pancreatic cancer [J]. Int J Mol Sci , 2023 , 24 ( 16 ): 12993 . DOI: 10.3390/ijms241612993 http://dx.doi.org/10.3390/ijms241612993 .
GU Z W , WANG L , DONG Q , et al . Aberrant LYZ expression in tumor cells serves as the potential biomarker and target for HCC and promotes tumor progression via csGRP78 [J]. Proc Natl Acad Sci USA , 2023 , 120 ( 29 ): e2215744120 . DOI: 10.1073/pnas.2215744120 http://dx.doi.org/10.1073/pnas.2215744120 .
LI C , GUAN R , LI W M , et al . Single-cell RNA sequencing reveals tumor immune microenvironment in human hypopharygeal squamous cell carcinoma and lymphatic metastasis [J]. Front Immunol , 2023 , 14 : 1168191 . DOI: 10.3389/fimmu.2023.1168191 http://dx.doi.org/10.3389/fimmu.2023.1168191 .
WU Z H , ZHANG Z X , LEI Z H , et al . CD14: biology and role in the pathogenesis of disease [J]. Cytokine Growth Factor Rev , 2019 , 48 : 24 - 31 . DOI: 10.1016/j.cytogfr.2019.06.003 http://dx.doi.org/10.1016/j.cytogfr.2019.06.003 .
WANG J P , HUANG J N , DING H , et al . Functional analysis of tumor-derived immunoglobulin lambda and its interacting proteins in cervical cancer [J]. BMC Cancer , 2023 , 23 ( 1 ): 929 . DOI: 10.1186/s12885-023-11426-9 http://dx.doi.org/10.1186/s12885-023-11426-9 .
FARROKHI YEKTA R , AREFI OSKOUIE A , REZAEI TAVIRANI M , et al . Decreased apolipoprotein A4 and increased complement component 3 as potential markers for papillary thyroid carcinoma: a proteomic study [J]. Int J Biol Markers , 2018 , 33 ( 4 ): 455 - 462 . DOI: 10.1177/1724600818787752 http://dx.doi.org/10.1177/1724600818787752 .
MA Y F , CHEN Y C , FANG D L , et al . The immune-related gene CD52 is a favorable biomarker for breast cancer prognosis [J]. Gland Surg , 2021 , 10 ( 2 ): 780 - 798 . DOI: 10.21037/gs-20-922 http://dx.doi.org/10.21037/gs-20-922 .
WU Z L , GE L Y , MA L L , et al . TPM2 attenuates progression of prostate cancer by blocking PDLIM7-mediated nuclear translocation of YAP1 [J]. Cell Biosci , 2023 , 13 ( 1 ): 39 . DOI: 10.1186/s13578-023-00993-w http://dx.doi.org/10.1186/s13578-023-00993-w .
PIAO X M , KANG H W , JEONG P , et al . A prognostic immune predictor, HLA-DRA, plays diverse roles in non-muscle invasive and muscle invasive bladder cancer [J]. Urol Oncol , 2021 , 39 ( 4 ): 237.e21 - 237.237 .e 29 . DOI: 10.1016/j.urolonc.2020.11.017 http://dx.doi.org/10.1016/j.urolonc.2020.11.017 .
KAMIZA A B , KAMIZA S , MATHEW C G . HLA-DRB1 alleles and cervical cancer: a meta-analysis of 36 case-control studies [J]. Cancer Epidemiol , 2020 , 67 : 101748 . DOI: 10.1016/j.canep.2020.101748 http://dx.doi.org/10.1016/j.canep.2020.101748 .
LI Y X , WANG P , YE D M , et al . IGHG1 induces EMT in gastric cancer cells by regulating TGF-β/SMAD3 signaling pathway [J]. J Cancer , 2021 , 12 ( 12 ): 3458 - 3467 . DOI: 10.7150/jca.56056 http://dx.doi.org/10.7150/jca.56056 .
TAVAZOIE M F , POLLACK I , TANQUECO R , et al . LXR/ApoE activation restricts innate immune suppression in cancer [J]. Cell , 2018 , 172 ( 4 ): 825 - 840.e18 . DOI: 10.1016/j.cell.2017.12.026 http://dx.doi.org/10.1016/j.cell.2017.12.026 .
BERGHAUS N , SCHREINER S , POOS A M , et al . Comparison of IGLV2-14 light chain sequences of patients with AL amyloidosis or multiple myeloma [J]. FEBS J , 2023 , 290 ( 17 ): 4256 - 4267 . DOI: 10.1111/febs.16805 http://dx.doi.org/10.1111/febs.16805 .
ABUZINADAH N , KUMAR POSA S , ALARFAJ A A , et al . Improved prediction of ovarian cancer using ensemble classifier and shaply explainable AI [J]. Cancers (Basel) , 2023 , 15 ( 24 ): 5793 . DOI: 10.3390/cancers15245793 http://dx.doi.org/10.3390/cancers15245793 .
0
Views
0
下载量
0
CSCD
Publicity Resources
Related Articles
Related Author
Related Institution
京公网安备11010802024621
