Modifications to "Deriving Chemosensitivity from Cell Lines"

This web page is designed to hold updates or changes to our web site supplement for the manuscript Deriving Chemosensitivity from Cell Lines: Forensic Bioinformatics and Reproducible Research in High-Throughput Biology by Keith A. Baggerly and Kevin R. Coombes. The main page is here.

A Starter Set

Since there's now been a lot of material posted, we've assembled a Starter Set for background.

The IOM Committee in Their Own Words

Since the IOM report was issued, there have been several presentations by members of the committee discussing the context of their recommendations. We have collected several of these on their own webpage.

The IOM Report, March 23, 2012

The full 274p report is available here. A "report brief" (5p) is here.

60 Minutes, Feb 12, 2012

The clip and transcript are available here.

Notes to the IOM re Moffitt Trial, Covered Nov 4, 2011

Documents sent to the IOM by Moffitt and the NCI pertaining to NCT00720096 have their own webpage.

Notes from the IOM Meeting Aug 22, 2011

MP3 audio files of the open session discussion with Duke representatives giving the institutional response to the situation have their own webpage.

Notes from the IOM Meeting Jun 30, 2011

MP3 audio files of the open session discussion with Dr. Lisa McShane, and some of the documents mentioned therein, have their own webpage.

Notes from the IOM Meeting Mar 30-31, 2011

These notes, which include our presentation and MP3 audio files of the open session talks, have their own webpage.

Our Annotation of NCI Documents Released to the IOM in Dec 2010

The Annotation
A Timeline
Simulations Described in the Annotation

Notes from the ENAR Session on Reproducibility, Mar 23, 2011

These notes, which include our presentation and WAV audio files of all discussions, have their own webpage.

Abbreviated Timeline of Events Since Publication

mid Sep, 2009: paper available online.
late Sep, 2009: Duke launches an internal investigation.
Oct 2, 2009: Story covered in The Cancer Letter.
Oct 6, 2009: Duke University suspends clinical trials NCT00509366 (contrasting cisplatin and pemetrexed) and NCT00564876 (contrasting pemetrexed and vinorelbine).
Oct 8, 2009: Moffitt Cancer Center terminates clinical trial NCT00720096 (contrasting doxorubicin and topotecan).
Oct 9, 2009: Suspensions covered in The Cancer Letter.
Oct 19, 2009: Duke University suspends clinical trial NCT00636441 (contrasting doxorubicin and docetaxel).
Oct 23, 2009: Lack of blinding in studies discussed in The Cancer Letter.
early Nov, 2009: New data for cisplatin and pemetrexed is posted at http://data.genome.duke.edu/JCO. Our analysis of this data is below the horizontal line.
Nov 9-10, 2009: Copies of our analysis are sent to the Duke investigation and to the NCI.
mid Nov, 2009: Main Duke data site http://data.genome.duke.edu/ and descendant pages "go dark".
early Jan, 2010: Story covered in The IMS Bulletin.
Jan 29, 2010: Duke announces end of internal investigation, and intent to restart clinical trials. Duke declines to make the investigation's report or the underlying data available. We make our November 2009 analysis public. This story is covered in The Cancer Letter.
early April, 2010: Main Duke data site http://data.genome.duke.edu/ and descendant pages reappear. Data files have been removed from http://data.genome.duke.edu/NatureMedicine.php and http://data.genome.duke.edu/JCO.php.
May 14, 2010: We comment on a redacted version of the reviewers' report obtained by The Cancer Letter under the Freedom of Information Act (FOIA).
July 16, 2010: The Cancer Letter reports that one of the trial PIs used to claim he was a Rhodes Scholar, but the Rhodes foundation disagrees.
July 19, 2010: A group of 33 biostatisticians and bioinformaticians write to Harold Varmus, the new head of the NCI, to ask that trials be suspended until the science is clarified.
July 16-20, 2010: Duke trials are resuspended.
Oct 22/29, 2010: The Cancer Letter reports that Nevins is calling for a retraction of the Hsu et al cisplatin/pemetrexed paper, citing problems identical to those we reported in Nov 2009.
Nov 9, 2010: Duke trials are terminated.
Nov 19, 2010: Retraction of Nature Medicine paper initiated.
Dec 20, 2010: First meeting of the new IOM committee to review the use of omics signatures in predicting patient outcomes. The NCI releases about 550 pages of documentation.
Jan 11, 2011: Nature news feature reports Duke deans and the acting head of the Duke IRB withheld our Nov 2009 report from the external reviewers.
Jan 14, 2011: We provide our annotation of the documents released by the NCI (see above).
Jan 28, 2011: The Cancer Letter reports the FDA is auditing the Duke trials.

Analysis of New Data for Cisplatin Posted in Nov 2009

In November 2009, new data in support of the paper by Hsu et al (JCO, 25:4350-7, 2007) was posted at http://data.genome.duke.edu/JCO. This was posted in the middle of an investigation into the validity of these models, and two years after clinical trials using these predictors were initiated. Our analysis is given below. These data included 59 "test samples" used to validate the performance of their predictors for cisplatin and pemetrexed. At least 43 of these samples were mislabeled. We say "at least" because the genes were so thoroughly mislabeled for the remaining 16 samples that we could not identify the sample of origin.