A comprehensive genomic pan-cancer classification using The Cancer Genome Atlas gene expression data

Abstract

BackgroundThe Cancer Genome Atlas (TCGA) has generated comprehensive molecular profiles. We aim to identify a set of genes whose expression patterns can distinguish diverse tumor types. Those features may serve as biomarkers for tumor diagnosis and drug development.MethodsUsing RNA-seq expression data, we undertook a pan-cancer classification of 9,096 TCGA tumor samples representing 31 tumor types. We randomly assigned 75% of samples into training and 25% into testing, proportionally allocating samples from each tumor type.ResultsWe could correctly classify more than 90% of the test set samples. Accuracies were high for all but three of the 31 tumor types, in particular, for READ (rectum adenocarcinoma) which was largely indistinguishable from COAD (colon adenocarcinoma). We also carried out pan-cancer classification, separately for males and females, on 23 sex non-specific tumor types (those unrelated to reproductive organs). Results from these gender-specific analyses largely recapitulated results when gender was ignored. Remarkably, more than 80% of the 100 most discriminative genes selected from each gender separately overlapped. Genes that were differentially expressed between genders included BNC1, FAT2, FOXA1, and HOXA11. FOXA1 has been shown to play a role for sexual dimorphism in liver cancer. The differentially discriminative genes we identified might be important for the gender differences in tumor incidence and survival.ConclusionsWe were able to identify many sets of 20 genes that could correctly classify more than 90% of the samples from 31 different tumor types using TCGA RNA-seq data. This accuracy is remarkable given the number of the tumor types and the total number of samples involved. We achieved similar results when we analyzed 23 non-sex-specific tumor types separately for males and females. We regard the frequency with which a gene appeared in those sets as measuring its importance for tumor classification. One third of the 50 most frequently appearing genes were pseudogenes; the degree of enrichment may be indicative of their importance in tumor classification. Lastly, we identified a few genes that might play a role in sexual dimorphism in certain cancers.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-017-3906-0) contains supplementary material, which is available to authorized users.

Deep Learning Based Tumor Type Classification Using Gene Expression Data

“…(2) Some CHOL samples are misclassified into LIHC due to the small number of CHOL samples. But using CNN do show an improvement in dealing with READ samples compared to the reference [11]. A comparison of accuracy as to each class is shown in table 3.…”

“…Binary classification (identification) of tumors has been found in some papers, but only paper [11] researched multiple tumor type classification using gene expression data. The authors applied GA/ KNN method to iteratively generate the subset of the genes (features) and then use KNN method to test the accuracy.…”

“…Also, the steps to generate heat map for each sample are also explained. Finally, in section 4, The accuracy of the tumor type classification is discussed and compared with the result of GA/KNN method in paper [11]. Also, top genes are picked from the heat-maps and validated by functional analysis.…”

AutoGenome: An AutoML Tool for Genomic Research

“…For example, the top 1 rank gene across all 24 cancer types is RPS27 (Fig. 3b), which has been observed highly expressed in various human cancers 55,56 ; the pseudogenes PA2G4P4 and H3F3C are reported to be functional in many cancers 57,58 (Fig. 3b); in cervical and endocervical cancers (CESC), a long non-coding RNA gene, MALAT1, ranks top 3 among all 8449 genes (Fig.…”

Transcriptomic biomarkers for prediction/classification of sample conditions in marine microeukaryotes (diatoms)

“…as "healthy" or "diseased" (Akbani et al 2015). The correlation of highly specific molecular readings, including transcript levels, with the biological state of test subjects allows the classification, diagnosis or corroboration of hypothetical or hidden conditions (Li et al 2017). This may also be applicable to biological and environmental questions, wherein measurable molecular biomarkers can indicate the biological and environmental status of study areas (Saito et al 2014).…”