Zhenglun Pan, of Shandong Provincial Hospital, China, and a multinational team of colleagues investigated studies that had been done on thirteen "gene-disease associations" (i.e., links between specific genes and certain diseases).

They first looked at the studies on these 13 gene-disease associations that were published in English language journals and that were logged into an international database of medical research called PubMed. They then searched a Chinese database for additional studies on these same 13 topics.

In the Chinese database, the researchers identified 161 Chinese studies on 12 of these gene-disease associations. Only 20 of these 161 studies were indexed in PubMed.

The researchers found important differences between Chinese and non-Chinese studies. For example, with one exception the first Chinese study appeared with a time lag (221 years) after the first non-Chinese study on the topic. The Chinese studies also showed significantly more prominent genetic effects than the non-Chinese studies. This was probably due to bias favoring the dissemination of impressive results, and this bias seemed to operate also beyond the Chinese literature.

Zhenglun Pan and colleagues' study is important because it suggests that the evidence on whether a particular gene is associated with a particular disease will be skewed, depending on whether you look only at English language studies or whether you also look at non-English language studies. The evidence would also be skewed depending on the extent of bias in favor of publishing "positive" spectacular results, while "negative" studies that are equally well done are not published.

"If all investigators working on the genetics of a specific disease were registered in a common network," say
'"/>

Contact: Paul Ocampo
pocampo@plos.org
415-624-1224
Public Library of Science
22-Nov-2005