Saturday, December 17, 2005

Should reviewers at Science have caught the errors in the Korean cloning work?

from Nicholas Wade in the New York Times:

"Should reviewers have caught some of this? Yeah, probably they should have," said John Gearhart, a stem cell expert at Johns Hopkins University and a member of Science's board of reviewers. "Obviously great claims require great proof, and maybe more people should review such a paper," he said.

Dr. Arthur Levine, dean of the University of Pittsburgh medical school, where a committee is investigating the work in the Science article for possible misconduct, said he agreed that Science's referees "might have been more critical, but that is hindsight." He added, "Almost six months have elapsed since the paper was published, and it had been widely read by many fine scientists without challenge."

Unlike readers, a journal's reviewers can demand more data on points that do not persuade them. But reviewers are unpaid, and their main task is to judge whether the data presented to them support the claim being made. Dr. Kennedy said that the reviewers could not be expected to detect deliberate falsehoods and that he could not see any generic fault in the peer review system.

Dr. Levine said he saw "terribly important lessons" in the Hwang incident, chiefly that the senior author of an article is responsible for its integrity and must therefore be "intimately familiar" with the data. However, the true test of science is replication of a claim by others, Dr. Levine said.

Other scientists have expressed the fear that a replication - the success by a second laboratory in cloning human cells - might have seemed to vindicate Dr. Hwang's work, if the present criticisms had not come to light. Any Nobel Prize might then have gone to him, not to the scientists who had apparently come second. "If the procedure works indeed and other labs would have repeated it, the credit would have gone to Hwang," Dr. Jaenisch said.

DNA fingerprint tests were indeed demanded by Science's reviewers for the 11 colonies allegedly derived from patients in the June 2005 paper, and Dr. Hwang provided them.

But critics have said that in some cases the fingerprints Dr. Hwang provided have the identical background noise, as if a single test were being presented twice, instead of two tests having been made independently. [Note: exactly the same problem as in the earlier fraudulent work of Jan-Hendrik Schon.]

Dr. Lanza also pointed to a similar article by Dr. Hwang in the August 2005 issue of the journal Molecular Reproduction and Development. The identical photograph, a test of pig embryonic cell colonies, seemed to have been presented twice but as representing different cells. The first photo shows embryonic cells said to have been cloned from an adult pig, and the second embryonic cells labeled as from a fertilized pig egg.
[Schon issue, again.]

Dr. Lanza also expressed doubt about the cloned dog, Snuppy, which Dr. Hwang announced in Nature, Science's rival publication, in August 2005. Nature's editors should have required a test of the mitochondrial DNA of Snuppy and the dog it was cloned from, Dr. Lanza said. If a true clone, the mitochondrial DNA of Snippy and the egg's donor would be different.
[that is, in this method, if a nucleus from a first dog's adult cell is inserted into an egg from a second dog, the mitochondrial DNA will be different, because the mitochondrial DNA will be from the second dog, not from the first dog (the one that is being cloned). The nuclear DNA of the clone (Snippy) will be identical to the nuclear DNA of the adult cell]

But Dr. Hwang, in a very brief report with little data, reported only that the nuclear DNA of Snuppy and the adult cell donor were identical. This would not rule out the possibility that the two dogs were in fact identical twins obtained by splitting an embryo and delaying the gestation of the second twin. "I think it's essential that he immediately allow independent testing of the original and cloned animals," Dr. Lanza said.

Dr. Gearhart agreed, saying, "That would be the very first thing that anyone would have asked."

[from the New York Times, downloaded on December 17, 2005.]

Although work that has appeared in both Science and Nature is being questioned, in part because of the apparent use of one photograph to represent more than one distinct thing, neither Nicholas Wade nor writers in other significant newspapers (e.g., Washington Post, LA Times) have brought up the obvious parallel to the fraudulent work by Jan-Hendrik Schon, which also appeared both in Science and Nature. Therein, Schon used one graph to represent more than one distinct thing. Such a repetition is something that a careful reviewer should catch. In the Schon case, once people were tipped off that this occurred, it was relatively easy to prove.

Separately, as to retraction policy at Science, Dr. Hwang Woo Suk has asked for retraction. In the Schon matter, Jan-Hendrik Schon did not ask for retraction, but the papers were finally retracted at Science, and later at Nature. In the Schon matter, US patent applications were withdrawn immediately at the conclusion of the study by Malcolm (Mac) Beasley.

In a different area, as to patent quality at the USPTO (relating to issues raised by Quillen and Webster, Lemley and Moore), one notes that the best scientific minds of the reviewers of Science failed to catch errors not only in the cloning work of Dr. Hwang but also in the solid state work of Jan-Hendrik Schon. In only a few years, the reviewers of Science have arguably passed by more "bad science" than the examiners at the USPTO.


Post a Comment

<< Home