Cancer Informatics最新文献

{"title":"A BIRC5^High COD1^Low Cancer Tissue Phenotype Indicates Poorer Prognosis of Metastatic Breast Cancer Patients.","authors":"Yujie Bai,&nbsp;Feng Yuan,&nbsp;Jing Yu,&nbsp;Yibei Si,&nbsp;Yiwen Zheng,&nbsp;Dongqing Li","doi":"10.1177/11769351221096655","DOIUrl":"https://doi.org/10.1177/11769351221096655","url":null,"abstract":"<p><p>Extensive data research is helpful to find sensitive biomarkers for prognostic prediction of metastatic breast cancer. Through analyzing multiple GEO datasets, literature retrieval, and verified in GEPIA datasets, we identify BIRC5 (Baculoviral IAP repeat containing 5) and CDO1 (Cysteine dioxygenase type 1) as DEGs (differentially expressed genes) between breast tumor and normal tissue and DEGs between metastatic breast cancer and breast cancer in situ. Then, we performed a series of in silico studies on BIRC5 and CDO1 using online tools including the UALCAN, TIMER, TCGA-BRCA, LinkedOmics Kaplan-Meier Plotter, and an R script for analysis. To verify the association of 2 genes expression and patients' clinical data, we detected BIRC5 and CDO1 mRNA in the tissue of 48 breast cancer patients. The results showed the tumor with BIRC5^high CDO1^low expression generally indicated patients' shorter overall (OS) and relapse-free survival (RFS). Specifically, BIRC5 and CDO1 levels significantly affect OS or RFS in patients with Lymph node metastasis and molecular subtypes of TNBC (triple-negative breast cancer) and Luminal A. A BIRC5^high tumor displayed a purer tumor purity and expressed more KIR receptors on NK cells while activating more FOXP3⁺CD25⁺ Treg cells. The CDO1^low tumors infiltrated with more immunocytes leading to less tumor purity. In our verified experiment, BIRC5 mRNA level in patients with stage III and over was significantly higher than in patients with stage 0 to II, but there were no significant differences among molecular subtyping groups; TNBC tissue expressed lower CDO1 mRNA level than HER2⁺ and Luminal type cancer tissue. In conclusion, a BIRC5^high CDO1^low expression type in breast cancer tissue indicates a poorer prognosis of patients. The potential mechanism might be increased BIRC5 expression in cancer tissue is likely to accompany NK cells inhibition, activating more Treg cells, and lacking effective CD8⁺ T cells proliferation. Meanwhile, CDO1 level is positively related to more immunocytes infiltration.</p>","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":" ","pages":"11769351221096655"},"PeriodicalIF":2.0,"publicationDate":"2022-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/67/30/10.1177_11769351221096655.PMC9208035.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"40237077","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An 8-Gene Signature for Classifying Major Subtypes of Non-Small-Cell Lung Cancer.","authors":"Mehdi Hamaneh,&nbsp;Yi-Kuo Yu","doi":"10.1177/11769351221100718","DOIUrl":"https://doi.org/10.1177/11769351221100718","url":null,"abstract":"<p><strong>Motivation: </strong>The precise diagnosis of the major subtypes, lung adenocarcinoma and lung squamous cell carcinoma, of non-small-cell lung cancer is of practical importance as some treatments are subtype-specific. However, in some cases diagnosis via the commonly-used method, that is staining the specimen using immunohistochemical markers, may be challenging. Hence, having a computational method that complements the diagnosis is desirable. In this paper, we propose a gene signature for this purpose.</p><p><strong>Results: </strong>We developed an expression-based method that systematically suggests a huge set of candidate gene signatures and finds the best candidate. By applying this method to a training set, the optimal gene signature was found by considering close to 765 billion candidate signatures. The 8-gene signature found for classifying the 2 aforementioned subtypes comprises TP63, CALML3, KRT5, PKP1, TESC, SPINK1, C9orf152, and KRT7. The signature achieved a high overall prediction accuracy of 0.936 when tested using 34 independent gene expression datasets obtained using different technologies and comprising 2556 adenocarcinoma and 1630 squamous cell carcinoma samples. Additionally, the signature performed well in clinically challenging cases, that is poorly differentiated tumors and specimens obtained from biopsies. In comparison with 2 previously reported signatures, our signature performed better in terms of overall accuracy and especially accuracy of classifying lung squamous cell carcinoma.</p><p><strong>Conclusions: </strong>Our signature is easy to use and accurate regardless of the technology used to obtain the gene expression profiles. It performs well even in clinically challenging cases and thus can assist pathologists in diagnosis of the ambiguous cases.</p>","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":" ","pages":"11769351221100718"},"PeriodicalIF":2.0,"publicationDate":"2022-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/b9/ce/10.1177_11769351221100718.PMC9201361.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"40042209","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Semi-Automated Computational Assessment of Cancer Organoid Viability Using Rapid Live-Cell Microscopy.","authors":"Joseph D Buehler, Cylaina E Bird, Milan R Savani, Lauren C Gattie, William H Hicks, Michael M Levitt, Mohamad El Shami, Kimmo J Hatanpaa, Timothy E Richardson, Samuel K McBrayer, Kalil G Abdullah","doi":"10.1177/11769351221100754","DOIUrl":"10.1177/11769351221100754","url":null,"abstract":"<p><p>The creation of patient-derived cancer organoids represents a key advance in preclinical modeling and has recently been applied to a variety of human solid tumor types. However, conventional methods used to assess in vivo tumor tissue treatment response are poorly suited for the evaluation of cancer organoids because they are time-intensive and involve tissue destruction. To address this issue, we established a suite of 3-dimensional patient-derived glioma organoids, treated them with chemoradiotherapy, stained organoids with non-toxic cell dyes, and imaged them using a rapid laser scanning confocal microscopy method termed \"Apex Imaging.\" We then developed and tested a fragmentation algorithm to quantify heterogeneity in the topography of the organoids as a potential surrogate marker of viability. This algorithm, SSDquant, provides a 3-dimensional visual representation of the organoid surface and a numerical measurement of the sum-squared distance (SSD) from the derived mass center of the organoid. We tested whether SSD scores correlate with traditional immunohistochemistry-derived cell viability markers (cellularity and cleaved caspase 3 expression) and observed statistically significant associations between them using linear regression analysis. Our work describes a quantitative, non-invasive approach for the serial measurement of patient-derived cancer organoid viability, thus opening new avenues for the application of these models to studies of cancer biology and therapy.</p>","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":"21 1","pages":"11769351221100754"},"PeriodicalIF":2.0,"publicationDate":"2022-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9150230/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"44955619","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Lift the Veil of Breast Cancers Using 4 or Fewer Critical Genes.","authors":"Zhengjun Zhang","doi":"10.1177/11769351221076360","DOIUrl":"https://doi.org/10.1177/11769351221076360","url":null,"abstract":"<p><p>Known genes in the breast cancer study literature could not be confirmed whether they are vital to breast cancer formations due to lack of convincing accuracy, although they may be biologically directly related to breast cancer based on present biological knowledge. It is hoped vital genes can be identified with the highest possible accuracy, for example, 100% accuracy and convincing causal patterns beyond what has been known in breast cancer. One hope is that finding gene-gene interaction signatures and functional effects may solve the puzzle. This research uses a recently developed competing linear factor analysis method in differentially expressed gene detection to advance the study of breast cancer formation. Surprisingly, 3 genes are detected to be differentially expressed in TNBC and non-TNBC (Her2, Luminal A, Luminal B) samples with 100% sensitivity and 100% specificity in 1 study of triple-negative breast cancers (TNBC, with 54 675 genes and 265 samples). These 3 genes show a clear signature pattern of how TNBC patients can be grouped. For another TNBC study (with 54 673 genes and 66 samples), 4 genes bring the same accuracy of 100% sensitivity and 100% specificity. Four genes are found to have the same accuracy of 100% sensitivity and 100% specificity in 1 breast cancer study (with 54 675 genes and 121 samples), and the same 4 genes bring an accuracy of 100% sensitivity and 96.5% specificity in the fourth breast cancer study (with 60 483 genes and 1217 samples). These results show the 4-gene-based classifiers are robust and accurate. The detected genes naturally classify patients into subtypes, for example, 7 subtypes. These findings demonstrate the clearest gene-gene interaction patterns and functional effects with the smallest numbers of genes and the highest accuracy compared with findings reported in the literature. The 4 genes are considered to be essential for breast cancer studies and practice. They can provide focused, targeted researches and precision medicine for each subtype of breast cancer. New breast cancer disease types may be detected using the classified subtypes, and hence new effective therapies can be developed.</p>","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":" ","pages":"11769351221076360"},"PeriodicalIF":2.0,"publicationDate":"2022-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/ab/27/10.1177_11769351221076360.PMC8851495.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"39800798","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Liquid Biopsy: A Distinctive Approach to the Diagnosis and Prognosis of Cancer.","authors":"Tetsuyuki Hirahata,&nbsp;Reeshan Ul Quraish,&nbsp;Afraz Ul Quraish,&nbsp;Shahan Ul Quraish,&nbsp;Munazzah Naz,&nbsp;Mohammad Abdul Razzaq","doi":"10.1177/11769351221076062","DOIUrl":"https://doi.org/10.1177/11769351221076062","url":null,"abstract":"Cancer is a leading cause of morbidity and mortality worldwide. Over the past decades, the concept of precision cancer medicine has emerged as a novel approach in the field of oncology that aims to tailor the most effective treatment options to each individual cancer patient based on the genetic profile of the tumor of each individual patient. Recently, tissue biopsy has become an essential part of cancer care and is widely used to characterize the tumor. However, tissue biopsy techniques face different challenges due to their invasiveness, cost, time, and adversity in potential sampling due to tissue heterogeneity. To overcome these issues, a non-invasive approach has developed, which is known as liquid biopsy. It is a simple, fast, and worthwhile technique based on the analysis of circulating tumor DNA (which is a fraction of cfDNA), circulating tumor cells (CTCs), and other tumor-derived material in blood plasma. This review provides an overview of the concept of liquid biopsy and briefly discusses the role of ctDNA and CTC analysis as tools for early diagnosis and prognosis of cancer. In this review, we also speculate on the advantages of liquid biopsy as opposed to tissue biopsy and postulate that liquid biopsy may be a comprehensive approach to overcome the current limitations associated with costly, invasive, and time-consuming tissue biopsy.","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":" ","pages":"11769351221076062"},"PeriodicalIF":2.0,"publicationDate":"2022-02-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/1f/d8/10.1177_11769351221076062.PMC8832574.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"39914505","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Chronotherapy of Early Colon Cancer: Advantage of Morning Dose Schedules.","authors":"David E Axelrod","doi":"10.1177/11769351211067697","DOIUrl":"https://doi.org/10.1177/11769351211067697","url":null,"abstract":"<p><p>Colon adenomas with proliferating mutant cells may progress to invasive carcinomas. Proliferation of cells in human colorectal tissue is circadian, greater in the interval 4 to 12 hours after midnight than 16 to 24 hours after midnight. We have tested the hypothesis that chemotherapy administered during the time of greater cell proliferation will be more effective than chemotherapy administered during the time of lesser proliferation. An agent-based computer model of cell proliferation in colon crypts was calibrated with measurements of cell numbers in human biopsy specimens. It was used to simulate cytotoxic chemotherapy of an early stage of colon cancer, adenomas with about 20% of mutant cells. Chemotherapy doses were scheduled at different 4-hour intervals during the 24-hour day, and repeated at weekly intervals. Chemotherapy administered at 4 to 8 hours after midnight cured mutant cells in 100% of 50 trials with an average time to cure of 7.82 days (s.e.m. = 0.99). In contrast, chemotherapy administered at 20 to 24 hours after midnight cured only 18% of 50 trials, with the average time to cure of 23.51 days (s.e.m. = 2.42). These simulation results suggest that clinical chemotherapy of early colon cancer may be more effective when given in the morning than later in the day.</p>","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":" ","pages":"11769351211067697"},"PeriodicalIF":2.0,"publicationDate":"2022-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/e8/aa/10.1177_11769351211067697.PMC8801641.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"39882095","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Risk Factors of White Blood Cell Progression Among Patients With Chronic Lymphocytic Leukemia at Felege Hiwot Referral Hospital, Bahir Dar, Ethiopia.","authors":"Gedam Derbew Addisia,&nbsp;Awoke Seyoum Tegegne,&nbsp;Denekew Bitew Belay,&nbsp;Mitiku Wale Muluneh,&nbsp;Mahider Abere Kassaw","doi":"10.1177/11769351211069902","DOIUrl":"https://doi.org/10.1177/11769351211069902","url":null,"abstract":"<p><strong>Background: </strong>Leukemia is a type of cancers that start in the bone marrow and produce a serious number of abnormal white blood cells. Bleeding and bruising problems, fatigue, fever, and an increased risk of infection are among symptoms of the disease. The main objective of this study is to identify the determinant of the progression rate of white blood cells among patients with chronic lymphocytic leukemia at Felege Hiwot Referral Hospital (FHRH), Bahir Dar, Ethiopia.</p><p><strong>Methods: </strong>A retrospective study design was conducted on 312 patients with chronic lymphocytic leukemia at FHRH, Bahir Dar, Ethiopia under treatment from 1 January 2017 to 31 December 2019. A linear mixed-effects model was considered for the progression of the white blood cell data.</p><p><strong>Results: </strong>The estimated coefficient of the fixed effect intercept was 84.68, indicating that the average white blood cell (WBC) count of the patients was 84.68 at baseline time by excluding all covariates in the model (<i>P</i>-value <.001). Male sex (<i>β</i> = 2.92, 95% confidence interval [CI] 0.58, 0.5.25), age (<i>β</i> = .17, 95% CI 0.08, 0.28), widowed/divorced marital status (<i>β</i> = 3.30, 95% CI 0.03, 6.57), medium chronic lymphocytic leukemia (CLL) stage (<i>β</i> = -4.34, 95% CI -6.57, -2.68), high CLL stage (<i>β</i> = -2.76, 95% CI -4.86, -0.67), hemoglobin (<i>β</i> = .15, 95% CI 0.07, 0.22), platelet (<i>β</i> = .09, 95% CI 0.02, 0.17), lymphocytes (<i>β</i> = .16, 95% CI 0.03, 0.29), red blood cell (RBC) (<i>β</i> = .17, 95% CI 0.09, 0.25), and follow-up time (<i>β</i> = .27, 95% CI 0.19, 0.36) were significantly associated with the average WBC count of chronic lymphocytic leukemia patients.</p><p><strong>Conclusions: </strong>The finding showed that age, sex, lymphocytic, stage of chronic lymphocytic leukemia, marital status, platelet, hemoglobin, RBC, and follow-up time were significantly associated with the average WBC count of chronic lymphocytic leukemia patients. Therefore, health care providers should give due attention and prioritize those identified factors and give frequent counseling about improving the health of chronic lymphocytic leukemia patients.</p>","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":" ","pages":"11769351211069902"},"PeriodicalIF":2.0,"publicationDate":"2022-01-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/3b/1c/10.1177_11769351211069902.PMC8771732.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"39728203","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Diagnostic Value of hTERT mRNA and in Combination With AFP, AFP-L3%, Des-γ-carboxyprothrombin for Screening of Hepatocellular Carcinoma in Liver Cirrhosis Patients HBV or HCV-Related","authors":"Hoang Bac Nguyen, Xuan Thi Thanh Le, Huy Huu Nguyen, Thanh Thanh Vo, M. Le, Ngan Nguyen, Thien Minh Do-Nguyen, Cong Minh Truong-Nguyen, Bang Suong Thi Nguyen","doi":"10.1177/11769351221100730","DOIUrl":"https://doi.org/10.1177/11769351221100730","url":null,"abstract":"Diagnosis of hepatocellular carcinoma (HCC) in early-stage, to give an effective treatment option and improve quality of life for cancer patients, is an important medical mission globally. Combination of AFP with some biomarkers may be more supportive in both diagnosis and screening of HCC, but the range value of these markers can be applied as daily markers were unclearly. In some studies, human telomerase reverse transcriptase (hTERT mRNA) was reported as an advantage marker to diagnose cancer. The present study identified serum of 340 patients that were infected chronic hepatitis B virus or hepatitis C virus and divided in 2 groups including Hepatocellular carcinoma (HCC) and liver cirrhosis (LC) to measure their values of hTERT mRNA, AFP, AFP-L3%, and DCP, as well as combination of them. As a result, the concentration of hTERT mRNA, AFP, AFP-L3%, and DCP in HCC groups were significantly higher than that in LC group (P < .01). For detecting HCC, hTERT mRNA had sensitivity of 88% and specificity of 96% (at the cutoff value of 31.5 copies/mL), AFP sensitivity of 73% and specificity of 92% (at the cutoff value of 5.1 ng/mL), AFP-L3% sensitivity of 69% and specificity of 90% (at the cutoff value of 1.05%), DCP sensitivity of 82% and specificity of 92% (at the cutoff value of 29.01 mAU/mL). The largest area under the curve (AUC) of combination hTERT mRNA with DCP was 0.932 (sensitivity of 98.2% and specificity of 88.2%). New combination of DCP with hTERT mRNA gave a useful choice for screening of HCC in chronic HBV or HCV patients associated liver cirrhosis.","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46634427","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A 25 Immune-Related Gene Pair Signature Predicts Overall Survival in Cervical Cancer","authors":"Huaqiu Chen, Huanyu Xie, Pengyu Wang, S. Yan, Yuanyuan Zhang, Guangming Wang","doi":"10.1177/11769351221090921","DOIUrl":"https://doi.org/10.1177/11769351221090921","url":null,"abstract":"Mounting evidence suggests that the tumor microenvironment plays an important role in the occurrence and development of cancer, with immune system dysfunction being closely related to malignant cancers. We aimed to screen immune-related genes (IRGs) to generate an IRG pair (IRGP)-based prognostic signature for cervical cancer (CC). Datasets were obtained from The Cancer Genome Atlas and Gene Expression Omnibus databases and used as training and validation cohorts, respectively. Using the ImmPort database, IRGs in control and CC samples were compared, and differentially expressed genes were identified to construct an IRGP prognostic signature. Based on this analysis, 25 IRGPs were identified as important factors for the prognosis of CC. Univariate and multivariate Cox regression analyses further showed that the IRGP signature was an independent prognostic factor of overall survival. In summary, we successfully constructed an IRGP prognostic signature of CC, providing insights into immunotherapy for CC.","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":" ","pages":""},"PeriodicalIF":2.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"47389905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TULIP: An RNA-seq-based Primary Tumor Type Prediction Tool Using Convolutional Neural Networks.","authors":"Sara Jones,&nbsp;Matthew Beyers,&nbsp;Maulik Shukla,&nbsp;Fangfang Xia,&nbsp;Thomas Brettin,&nbsp;Rick Stevens,&nbsp;M Ryan Weil,&nbsp;Satishkumar Ranganathan Ganakammal","doi":"10.1177/11769351221139491","DOIUrl":"https://doi.org/10.1177/11769351221139491","url":null,"abstract":"<p><strong>Background: </strong>With cancer as one of the leading causes of death worldwide, accurate primary tumor type prediction is critical in identifying genetic factors that can inhibit or slow tumor progression. There have been efforts to categorize primary tumor types with gene expression data using machine learning, and more recently with deep learning, in the last several years.</p><p><strong>Methods: </strong>In this paper, we developed four 1-dimensional (1D) Convolutional Neural Network (CNN) models to classify RNA-seq count data as one of 17 highly represented primary tumor types or 32 primary tumor types regardless of imbalanced representation. Additionally, we adapted the models to take as input either all Ensembl genes (60,483) or protein coding genes only (19,758). Unlike previous work, we avoided selection bias by not filtering genes based on expression values. RNA-seq count data expressed as FPKM-UQ of 9,025 and 10,940 samples from The Cancer Genome Atlas (TCGA) were downloaded from the Genomic Data Commons (GDC) corresponding to 17 and 32 primary tumor types respectively for training and validating the models.</p><p><strong>Results: </strong>All 4 1D-CNN models had an overall accuracy of 94.7% to 97.6% on the test dataset. Further evaluation indicates that the models with protein coding genes only as features performed with better accuracy compared to the models with all Ensembl genes for both 17 and 32 primary tumor types. For all models, the accuracy by primary tumor type was above 80% for most primary tumor types.</p><p><strong>Conclusions: </strong>We packaged all 4 models as a Python-based deep learning classification tool called TULIP (TUmor CLassIfication Predictor) for performing quality control on primary tumor samples and characterizing cancer samples of unknown tumor type. Further optimization of the models is needed to improve the accuracy of certain primary tumor types.</p>","PeriodicalId":35418,"journal":{"name":"Cancer Informatics","volume":"21 ","pages":"11769351221139491"},"PeriodicalIF":2.0,"publicationDate":"2022-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ftp.ncbi.nlm.nih.gov/pub/pmc/oa_pdf/b1/0b/10.1177_11769351221139491.PMC9729992.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"10333449","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}