==== Front Genes Dis Genes Dis Genes & Diseases 2352-4820 2352-3042 Chongqing Medical University S2352-3042(23)00005-3 10.1016/j.gendis.2022.12.010 Correspondence False discovery rate control in cancer biomarker selection Li Zhaoming fcclizm@zzu.edu.cn Department of Oncology, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, Henan 450052, China 20 1 2023 7 2023 20 1 2023 10 4 11411142 17 10 2022 27 12 2022 © 2023 The Authors. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co., Ltd. 2023 Chongqing Medical University https://creativecommons.org/licenses/by/4.0/ This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). ==== Body pmcWe read with great interest the manuscript “LPCAT1 functions as a novel prognostic molecular marker in hepatocellular carcinoma” by Zhang et al in a recent issue of Genes & Diseases.1 The authors conducted bioinformatics analyses using high throughput RNA sequencing data from TCGA to demonstrate that LPCAT1 is a novel and effective prognostic marker for hepatocellular carcinoma. We appreciate the contributions of the authors on the subject, nonetheless, we have some concerns that should be clarified in the following issues. The standard P-value was invented for testing individual hypotheses. There is an obvious problem when analyzing gene expression data collected via sequencing of multiple genomes, as this usually involves testing from several thousands to tens of thousands of hypotheses simultaneously. In genome sequencing studies most researchers are keenly aware of the potentially high rate of false positives and the need to control it. One key statistical shift is the move away from the well-known P-value to false discovery rate (FDR).2,3 The FDR of a test is defined as the expected proportion of false positives among the declared significant results.3, 4, 5 Because of this directly useful interpretation, FDR is a more convenient scale to work on instead of the P-value scale. However, in Zhang's report, the multiple test correction was not applied to the P values shown in Figure 2A and 7 and stated in the text. This seems to be required since the authors tested the association between each gene and outcome individually. The correlation between gene expression and overall survival in the same hepatocellular carcinoma RNA sequencing data from TCGA were re-analyzed by the Genomics Analysis and Visualization Platform (http://r2.amc.nl) and the results were corrected for multiple gene testing by FDR. The potential LPCAT1-related tumor genes reported by the author and their adjusted P-value were provided in Table 1. It showed that the expression of several genes (CCNB2, CENPF, and UBE2C) had no associations with overall survival, which is different from Figure 7 in the report.Table 1 The association of LPCAT1-related tumor genes with overall survival. Results were corrected for multiple gene testing by false discovery rate. Table 1No. Gene Probeset Adjusted P-value 1 CDC20 CDC20_991 0.001464779 2 CDCA8 CDCA8_55143 0.005409055 3 LPCAT1 LPCAT1_79888 0.00569335 4 TPX2 TPX2_22974 0.00691434 5 DLGAP5 DLGAP5_9787 0.008015183 6 CCNB1 CCNB1_891 0.008149167 7 MAD2L1 MAD2L1_4085 0.009098616 8 KIF4A KIF4A_24137 0.00910015 9 NUF2 NUF2_83540 0.009505066 10 CENPA CENPA_1058 0.011305291 11 KIF11 KIF11_3832 0.011942813 12 KIF20A KIF20A_10112 0.012491643 13 BIRC5 BIRC5_332 0.013002953 14 KIF2C KIF2C_11004 0.013481529 15 CDK1 CDK1_983 0.015119904 16 TTK TTK_7272 0.018113202 17 PLK1 PLK1_5347 0.018869319 18 BUB1 BUB1_699 0.023996996 19 BUB1B BUB1B_701 0.025563704 20 RRM2 RRM2_6241 0.025577274 21 NCAPG NCAPG_64151 0.029011476 22 TOP2A TOP2A_7153 0.029439732 23 NDC80 NDC80_10403 0.029498546 24 RACGAP1 RACGAP1_29127 0.030345955 25 KIF18A KIF18A_81930 0.035375811 26 CEP55 CEP55_55165 0.036683405 27 CENPE CENPE_1062 0.041494437 28 AURKB AURKB_9212 0.047951168 29 CCNB2 CCNB2_9133 not significant 30 CENPF CENPF_1063 not significant 31 UBE2C UBE2C_11065 not significant Conflict of interests The author declares no potential conflict of interests. Peer review under responsibility of Chongqing Medical University. ==== Refs References 1 Zhang H. Xu K. Xiang Q. LPCAT1 functions as a novel prognostic molecular marker in hepatocellular carcinoma Genes Dis 9 1 2022 151 164 35005115 2 Pawitan Y. Michiels S. Koscielny S. False discovery rate, sensitivity and sample size for microarray studies Bioinformatics 21 13 2005 3017 3024 15840707 3 Storey J.D. Tibshirani R. Statistical significance for genomewide studies Proc Natl Acad Sci U S A 100 16 2003 9440 9445 12883005 4 Shen A. Fu H. He K. False discovery rate control in cancer biomarker selection using knockoffs Cancers 11 6 2019 744 31146393 5 Benjamini Y. Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing J R Stat Soc Series B Stat Methodol 57 1 1995 289 300