Mixed up genes

A significant part of genetic studies may contain incorrectly labeled samples

Julia Korowski, XX2 century, based on EurekAlert: Researchers discover extensive mislabeling of gene expression samples

At least a third of studies on gene expression contain incorrectly labeled samples – this is the conclusion reached by scientists at the University of British Columbia (University of British Columbia). The results of the work are published on the F1000Research platform (Toker et al., Whose sample is it anyway? Widespread misannotation of samples in transcriptomics studies).

Re-analyzing publicly available data from four Parkinson's disease studies, Canadian scientists found that in two of them the expression of polo-specific genes does not coincide with the information about the subject's gender indicated by the geneticists themselves. Surprised researchers decided to find out how often such errors occur. To do this, they analyzed the expression of the genes of the polo-specific chromosomes in the same way as last time. Female individuals express some genes located on the X chromosome, while genes located on the Y chromosome express only males. To understand whether the labeling of the sample is correct, it is enough to compare the expression level of these genes with the gender "on the label".

The scientific team studied 70 data sets from 4043 samples. 46% of them contained at least one discrepancy between gene expression and the specified gender. Based on these data, scientists calculated that samples with incorrect labeling can be detected in at least 33% (or even 60%) of all studies on gene expression. The authors note that the methods they used do not allow to register cases when incorrect labeling is not related to gender – for example, if there was a mixing of samples of the material of two subjects of the same sex.

The scientific group tried to find out at what stage of the study such errors occur. Although in most cases the exact origin of the samples could not be determined, scientists found that incorrect labeling often took place before the data analysis began, and sometimes researchers confused the test tubes in the laboratory.

"Scientists have long known about the importance of sex markers for quality control, so it was surprising to find such obvious problems in so many studies," says Lilah Toker, a member of the scientific group. "We hope that our work will encourage other scientists to be more attentive." "Although these figures are alarming in themselves, the actual number of errors is likely to be higher, since gender analysis can only recognize a gender discrepancy and does not help to identify incorrect labeling of samples obtained as a result of case-control studies or samples with the same gender," notes the reviewer of the article Hans van Bokhoven (Hans van Bokhoven).

Incorrect labeling of samples can affect the reliability of the results and lead to the fact that the samples will be unsuitable for further use, therefore, the authors insist that a re-analysis of the expression of polo-specific genes should become a mandatory procedure.

Gene expression is a process in which the hereditary information of a gene is transformed into a functional product – RNA or protein. Some stages of gene expression can be regulated. Regulation of gene expression allows cells to control their own structure and function and is the basis of cell differentiation, morphogenesis and adaptation. Gene expression is a substrate for evolutionary changes, since control over the time, place, and quantitative characteristics of expression of one gene can have an impact on the functions of other genes in the whole organism.

Portal "Eternal youth" http://vechnayamolodost.ru  14.10.2016