Corrigendum to “Single cell sequencing reveals heterogeneity within ovarian cancer epithelium and cancer associated stromal cells” [Gyncol. Oncol. 144 (2017) 598–606] (Gynecologic Oncology (2017) 144(3) (598–606), (S0090825817300598) (10.1016/j.ygyno.2017.01.015))

Boris J Winterhoff, Makayla Maile, Amit Kumar Mitra, Attila Sebe, Martina Bazzaro, Melissa A Geller, Juan E Abrahante - Llorens, Molly E Klein, Raffaele Hellweg, Sally A Mullany, Kenneth Beckman, Jerry Daniel, Tim Starr

Research output: Contribution to journalComment/debate

Abstract

The authors regret that there is an error in their originally published article. The error in their single cell gene expression study affects many of the figures and tables presented in the manuscript entitled “Single cell sequencing reveals heterogeneity within ovarian cancer epithelium and cancer associated stromal cells” but does not affect the major conclusions presented in the manuscript. In the published article the authors presented their analysis of single cell RNA sequencing of 66 cells from a single patient. They analyzed single cell RNA expression data generated using next generation sequencing technologies including 10× Genomics and Illumina. The method for estimating RNA transcript abundance is based on read counts. Essentially, the more reads that map to a transcript, the more abundant the transcript. They used the RNA-Seq by Expectation Maximization (RSEM) method for analyzing the data [1]. The initial output of this method is “expected counts” (EC) which is basically the number of reads that map to a specific transcript for a specific sample. This number is then normalized for each sample based on the size of the gene and the total number of read counts attributed to that sample. This results in a normalized gene expression value, referred to as Transcripts Per Million (TPM). In the published manuscript they stated that they used TPM values as the basis for our analysis. They recently discovered that they had mistakenly used EC values, instead of TPM values for gene expression levels. The authors have re-analyzed the data using TPM values instead of the EC values. In general, all of the figures and tables are similar, except that a few of the cells originally classified in one group were placed in a different group, and the gene lists and the rank of genes presented in these lists changed. They have generated new figures and gene lists using the TPM expression values and present these new figures and tables in this corrigendum. The figures that have changed from the original manuscript include Figs. 1–5 Tables 1 & 2 Supplemental Figs. 1, 2, 3, 4, 5, 10, and 11, and Supplemental Tables 1–6. Supplemental Figs. 6, 7, 8, 9 and 12 were unaffected by this error. They also present a revised manuscript and a revised supplemental methods document with all of the changes included. The main changes to the manuscript include: 1) The average deviation cut-off value used to define highly variable genes was changed from 3.0 to 2.4, resulting in a list of 4272 highly expressed genes, with 399 of these genes defined as highly variable.2) Clustering using EC values placed 45 cells into the “epithelial group” and 21 cells into the “stromal group” while the re-analysis using TPM values re-classified 6 of the “epithelial” cells as “stromal” cells.3) Using TPM expression values, the “epithelial” cells clustered with the “Differentiated” molecular subtype, not the “Proliferative” molecular subtype.4) A few of the cells classified using our functional classification approach changed their classification grouping.The abstract and conclusions stated therein are unchanged in this new analysis. Based on their re-analysis they found that clustering results based on EC expression values are very similar to clustering results based on TPM expression values. The authors would like to apologise for any inconvenience caused.

Original languageEnglish (US)
Pages (from-to)182-186
Number of pages5
JournalGynecologic Oncology
Volume151
Issue number1
DOIs
StatePublished - Oct 1 2018

Fingerprint

Stromal Cells
Manuscripts
Ovarian Neoplasms
Epithelium
Ficus
Genes
Neoplasms
Cluster Analysis
Epithelial Cells
RNA
Gene Expression
Single-Cell Analysis
RNA Sequence Analysis
Genomics
corrigendum
Emotions
Technology

PubMed: MeSH publication types

  • Journal Article
  • Published Erratum

Cite this

@article{e5d07a0302a14ca7a8d5d6b362494b75,
title = "Corrigendum to “Single cell sequencing reveals heterogeneity within ovarian cancer epithelium and cancer associated stromal cells” [Gyncol. Oncol. 144 (2017) 598–606] (Gynecologic Oncology (2017) 144(3) (598–606), (S0090825817300598) (10.1016/j.ygyno.2017.01.015))",
abstract = "The authors regret that there is an error in their originally published article. The error in their single cell gene expression study affects many of the figures and tables presented in the manuscript entitled “Single cell sequencing reveals heterogeneity within ovarian cancer epithelium and cancer associated stromal cells” but does not affect the major conclusions presented in the manuscript. In the published article the authors presented their analysis of single cell RNA sequencing of 66 cells from a single patient. They analyzed single cell RNA expression data generated using next generation sequencing technologies including 10× Genomics and Illumina. The method for estimating RNA transcript abundance is based on read counts. Essentially, the more reads that map to a transcript, the more abundant the transcript. They used the RNA-Seq by Expectation Maximization (RSEM) method for analyzing the data [1]. The initial output of this method is “expected counts” (EC) which is basically the number of reads that map to a specific transcript for a specific sample. This number is then normalized for each sample based on the size of the gene and the total number of read counts attributed to that sample. This results in a normalized gene expression value, referred to as Transcripts Per Million (TPM). In the published manuscript they stated that they used TPM values as the basis for our analysis. They recently discovered that they had mistakenly used EC values, instead of TPM values for gene expression levels. The authors have re-analyzed the data using TPM values instead of the EC values. In general, all of the figures and tables are similar, except that a few of the cells originally classified in one group were placed in a different group, and the gene lists and the rank of genes presented in these lists changed. They have generated new figures and gene lists using the TPM expression values and present these new figures and tables in this corrigendum. The figures that have changed from the original manuscript include Figs. 1–5 Tables 1 & 2 Supplemental Figs. 1, 2, 3, 4, 5, 10, and 11, and Supplemental Tables 1–6. Supplemental Figs. 6, 7, 8, 9 and 12 were unaffected by this error. They also present a revised manuscript and a revised supplemental methods document with all of the changes included. The main changes to the manuscript include: 1) The average deviation cut-off value used to define highly variable genes was changed from 3.0 to 2.4, resulting in a list of 4272 highly expressed genes, with 399 of these genes defined as highly variable.2) Clustering using EC values placed 45 cells into the “epithelial group” and 21 cells into the “stromal group” while the re-analysis using TPM values re-classified 6 of the “epithelial” cells as “stromal” cells.3) Using TPM expression values, the “epithelial” cells clustered with the “Differentiated” molecular subtype, not the “Proliferative” molecular subtype.4) A few of the cells classified using our functional classification approach changed their classification grouping.The abstract and conclusions stated therein are unchanged in this new analysis. Based on their re-analysis they found that clustering results based on EC expression values are very similar to clustering results based on TPM expression values. The authors would like to apologise for any inconvenience caused.",
author = "Winterhoff, {Boris J} and Makayla Maile and Mitra, {Amit Kumar} and Attila Sebe and Martina Bazzaro and Geller, {Melissa A} and {Abrahante - Llorens}, {Juan E} and Klein, {Molly E} and Raffaele Hellweg and Mullany, {Sally A} and Kenneth Beckman and Jerry Daniel and Tim Starr",
year = "2018",
month = "10",
day = "1",
doi = "10.1016/j.ygyno.2018.07.015",
language = "English (US)",
volume = "151",
pages = "182--186",
journal = "Gynecologic Oncology",
issn = "0090-8258",
publisher = "Academic Press Inc.",
number = "1",

}

TY - JOUR

T1 - Corrigendum to “Single cell sequencing reveals heterogeneity within ovarian cancer epithelium and cancer associated stromal cells” [Gyncol. Oncol. 144 (2017) 598–606] (Gynecologic Oncology (2017) 144(3) (598–606), (S0090825817300598) (10.1016/j.ygyno.2017.01.015))

AU - Winterhoff, Boris J

AU - Maile, Makayla

AU - Mitra, Amit Kumar

AU - Sebe, Attila

AU - Bazzaro, Martina

AU - Geller, Melissa A

AU - Abrahante - Llorens, Juan E

AU - Klein, Molly E

AU - Hellweg, Raffaele

AU - Mullany, Sally A

AU - Beckman, Kenneth

AU - Daniel, Jerry

AU - Starr, Tim

PY - 2018/10/1

Y1 - 2018/10/1

N2 - The authors regret that there is an error in their originally published article. The error in their single cell gene expression study affects many of the figures and tables presented in the manuscript entitled “Single cell sequencing reveals heterogeneity within ovarian cancer epithelium and cancer associated stromal cells” but does not affect the major conclusions presented in the manuscript. In the published article the authors presented their analysis of single cell RNA sequencing of 66 cells from a single patient. They analyzed single cell RNA expression data generated using next generation sequencing technologies including 10× Genomics and Illumina. The method for estimating RNA transcript abundance is based on read counts. Essentially, the more reads that map to a transcript, the more abundant the transcript. They used the RNA-Seq by Expectation Maximization (RSEM) method for analyzing the data [1]. The initial output of this method is “expected counts” (EC) which is basically the number of reads that map to a specific transcript for a specific sample. This number is then normalized for each sample based on the size of the gene and the total number of read counts attributed to that sample. This results in a normalized gene expression value, referred to as Transcripts Per Million (TPM). In the published manuscript they stated that they used TPM values as the basis for our analysis. They recently discovered that they had mistakenly used EC values, instead of TPM values for gene expression levels. The authors have re-analyzed the data using TPM values instead of the EC values. In general, all of the figures and tables are similar, except that a few of the cells originally classified in one group were placed in a different group, and the gene lists and the rank of genes presented in these lists changed. They have generated new figures and gene lists using the TPM expression values and present these new figures and tables in this corrigendum. The figures that have changed from the original manuscript include Figs. 1–5 Tables 1 & 2 Supplemental Figs. 1, 2, 3, 4, 5, 10, and 11, and Supplemental Tables 1–6. Supplemental Figs. 6, 7, 8, 9 and 12 were unaffected by this error. They also present a revised manuscript and a revised supplemental methods document with all of the changes included. The main changes to the manuscript include: 1) The average deviation cut-off value used to define highly variable genes was changed from 3.0 to 2.4, resulting in a list of 4272 highly expressed genes, with 399 of these genes defined as highly variable.2) Clustering using EC values placed 45 cells into the “epithelial group” and 21 cells into the “stromal group” while the re-analysis using TPM values re-classified 6 of the “epithelial” cells as “stromal” cells.3) Using TPM expression values, the “epithelial” cells clustered with the “Differentiated” molecular subtype, not the “Proliferative” molecular subtype.4) A few of the cells classified using our functional classification approach changed their classification grouping.The abstract and conclusions stated therein are unchanged in this new analysis. Based on their re-analysis they found that clustering results based on EC expression values are very similar to clustering results based on TPM expression values. The authors would like to apologise for any inconvenience caused.

AB - The authors regret that there is an error in their originally published article. The error in their single cell gene expression study affects many of the figures and tables presented in the manuscript entitled “Single cell sequencing reveals heterogeneity within ovarian cancer epithelium and cancer associated stromal cells” but does not affect the major conclusions presented in the manuscript. In the published article the authors presented their analysis of single cell RNA sequencing of 66 cells from a single patient. They analyzed single cell RNA expression data generated using next generation sequencing technologies including 10× Genomics and Illumina. The method for estimating RNA transcript abundance is based on read counts. Essentially, the more reads that map to a transcript, the more abundant the transcript. They used the RNA-Seq by Expectation Maximization (RSEM) method for analyzing the data [1]. The initial output of this method is “expected counts” (EC) which is basically the number of reads that map to a specific transcript for a specific sample. This number is then normalized for each sample based on the size of the gene and the total number of read counts attributed to that sample. This results in a normalized gene expression value, referred to as Transcripts Per Million (TPM). In the published manuscript they stated that they used TPM values as the basis for our analysis. They recently discovered that they had mistakenly used EC values, instead of TPM values for gene expression levels. The authors have re-analyzed the data using TPM values instead of the EC values. In general, all of the figures and tables are similar, except that a few of the cells originally classified in one group were placed in a different group, and the gene lists and the rank of genes presented in these lists changed. They have generated new figures and gene lists using the TPM expression values and present these new figures and tables in this corrigendum. The figures that have changed from the original manuscript include Figs. 1–5 Tables 1 & 2 Supplemental Figs. 1, 2, 3, 4, 5, 10, and 11, and Supplemental Tables 1–6. Supplemental Figs. 6, 7, 8, 9 and 12 were unaffected by this error. They also present a revised manuscript and a revised supplemental methods document with all of the changes included. The main changes to the manuscript include: 1) The average deviation cut-off value used to define highly variable genes was changed from 3.0 to 2.4, resulting in a list of 4272 highly expressed genes, with 399 of these genes defined as highly variable.2) Clustering using EC values placed 45 cells into the “epithelial group” and 21 cells into the “stromal group” while the re-analysis using TPM values re-classified 6 of the “epithelial” cells as “stromal” cells.3) Using TPM expression values, the “epithelial” cells clustered with the “Differentiated” molecular subtype, not the “Proliferative” molecular subtype.4) A few of the cells classified using our functional classification approach changed their classification grouping.The abstract and conclusions stated therein are unchanged in this new analysis. Based on their re-analysis they found that clustering results based on EC expression values are very similar to clustering results based on TPM expression values. The authors would like to apologise for any inconvenience caused.

UR - http://www.scopus.com/inward/record.url?scp=85051037138&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051037138&partnerID=8YFLogxK

U2 - 10.1016/j.ygyno.2018.07.015

DO - 10.1016/j.ygyno.2018.07.015

M3 - Comment/debate

VL - 151

SP - 182

EP - 186

JO - Gynecologic Oncology

JF - Gynecologic Oncology

SN - 0090-8258

IS - 1

ER -