UniProt release 2011_03
Published March 8, 2011
Dealing with erroneous information: a tricky task
There are many reasons for mistakes in databases, including annotation errors. However, sometimes the annotation is correct, but the original source of information contains erroneous data. Histone arginine demethylase JMJD6 is a good example of the problems raised when this happens.
Histone arginine demethylase JMJD6 was discovered in 2000. At the origin of this discovery lies a monoclonal antibody (mAb217) raised against stimulated macrophages. Phosphatidylserine-displaying liposomes inhibited the binding of mAb217 to macrophages and the antibody prevented the uptake of apoptotic cells. These characteristics suggested that mAb217 interacted with a receptor for phosphatidylserine on the membrane. Using this antibody, a 48-kDa protein was isolated and called “Phosphatidylserine receptor” (PTDSR). The effect of the deletion of the corresponding gene was investigated in knockout mice, but the reactivity of mAb217 was not compared between cells from knockout and wild-type animals. When this experiment was finally carried out, the result was quite surprising: similar staining patterns were observed with cells of both genotypes. It appeared that mAb217 could bind weakly to a PTDSR peptide, but the antibody mainly recognizes another membrane-associated protein. Parallel studies revealed that PTDRS was actually a nuclear protein, an unlikely location for a membrane receptor. After several years and many publications in high-profile journals, it was eventually demonstrated that the 48-kDa protein is not a phosphatidylserine receptor, but a dioxygenase that acts in the nucleus as a histone arginine demethylase and a lysyl-hydroxylase. Since then, other genes have emerged as candidates for the role of phosphatidylserine receptor, including STAB2 , BAI1 and TIMD4.
Once the true function of the 48-kDa protein had been established, curators were faced with the challenge of updating the existing annotation to reflect this. First of all, the original recommended protein name “Phosphatidylserine receptor” had to be modified into “Bifunctional arginine demethylase and lysyl-hydroxylase JMJD6”. “Phosphatidylserine receptor” became an alternative name, as it is UniProtKB policy to keep all protein names, even obsolete ones, to facilitate the identification of the protein of interest. The problem in this case is that the obsolete name is misleading. In order to clarify this, the update of the ‘General annotation’ section included the addition of a ‘Caution’ comment, as well as the review of other subsections, such as ‘Function’ or ‘Subcellular location’. In the ‘Caution’ comment, the attention of the users is drawn to the ambiguity of the ‘Alternative name’, as well as to the erroneous conclusions reported in published references still cited in the entries. These references describe the original sequence of the protein and as such cannot be simply deleted from the entry, since this contribution has to be acknowledged. In addition, it confirms for the users that this information has been reviewed in the context of the protein and has not been overlooked.
We thus advise our users to carefully read the ‘Caution’ subsections found in entries which have had an interesting evolution (examples). Our users are also encouraged to send us feedback using the option “Contribute” at the top of each entry if they find mistakes or inconsistencies in our entries.
It is possible to track all changes occurring in an entry across releases by clicking on ‘History’, an option available at the top of each entry. For example, the major update of the human PTDSR entry can be visualized by comparing version 27 (which contained the original information) with the current one.
Changes concerning the controlled vocabulary for PTMsNew terms for the feature key ‘Cross-link’ (‘CROSSLNK’ in the flat file):
- Cyclopeptide (His-Asn)
- Cyclopeptide (His-Asp)
- Lysino-D-alanine (Lys)