I have rage.

Part of my PhD research is to characterize the digestibility of a set of Arabidopsis mutants. We chose these mutants because they were described in a Science paper that came out in June 2011. Now, Science papers are a big deal, even though they tend to be only a few pages long. There is great prestige associated with publishing in Science.

I wrote to the main author and asked him for homozygous seeds for the mutants and the transgenics they used in the paper. He said he would only send them to us if we would give him co-authorship on any paper we publish using these lines. This sort of request is absolutely not allowed. Once a plant line has been described, the seeds must be freely available to researchers. Nevertheless, Pat decided to say OK. Everyone I have told about this has agreed that this is wrong all around.

So the transgenic lines should be expressing green fluorescent protein in all their tissues. The presence of fluorescence is easily checked with fluorescence microscopy. I grew up some seeds and checked the plants and there was no fluorescence. I grew some more, checked them again - nothing, not in the roots, not in the leaves. These transgenic lines should be homozygous, if the researchers did their work properly, so all the plants I grow from the seed should be gaily fluorescing. All right. Whatever.

I go to the mutant lines. These lines are made by transforming the plants to have a chunk of extra DNA inserted somewhere in their genome. This chunk of DNA (T-DNA) disrupts whatever gene it is inserted into, so that gene is no longer expressed. There was a huge project in the early 2000s to find mutants with  T-DNA inserts in each gene in Arabidopsis, and for the most part it was very successful.

The transgenic plants were initially selected by growing the plants in the presence of antibiotics - only the transgenics could survive. When you order seeds from the seed bank, though, they come with a nice big warning that you cannot use antibiotic selection any more. You must find homozygous lines by PCR screening. Pat decides I should find homozygous lines with antibiotic screening, and I think she knows what she's talking about, so we try it. Eight months later, none of the plants from the seeds I have painstakingly made will grow in the presence of antibiotics. I ask Eric's postdoc Karl about this and he says, non non non, you must do PCR screening. Pat doesn't believe this could be true but somehow I convince her.

So now I have to design primers for the PCR screening. The people who published in Science included their PCR primers in the supplemental data, so I gather all those and start checking to see that they match the genomic sequence for each mutant. They do not. Several of the mutants, it turns out, have more than one insert - the first insert is in the gene of interest while the second is somewhere totally elsewhere in the genome. The primers that are published in the Science paper correspond to this other region of the genome. For other mutants, the primer sequence shows up more than once, which means the PCR will produce multiple products of different sizes and the results will be unclear. Their results, however, appear to be exceedingly clear, somehow.

Furthermore, it is best practice to make sure, somehow, that each of the mutants truly has only one T-DNA insertion, to be sure that any mutant phenotype you see is definitely due to a mutation in your gene. You can do this either by crossing the mutant to the wild-type plant and then selecting for lines that still have your mutation or by checking a different line that also has a T-DNA insertion in or around your gene of interest, which should still have the same mutant phenotype. Generally you must do one of these things to be able to publish. These people, who published in Science, did neither of these things.

And, the genes I am interested in are in a multi-gene family. There is a high chance that if one of the genes no longer functions due to the T-DNA insertion, a different member of the same family will pick up the slack so that there is no mutant phenotype. However, this group found the same mutant phenotype for every mutant line they examined! Unheard-of good luck!

I met the lead author at the Gordon Conference, and he had bad breath and was a very slimy character. I have found other little problems in this Science paper but now I don't trust anything that he says! Fortunately, I have a plan so that I won't have to deal with these dumb mutant lines any more, as long as Pat agrees to it. But I will have rage for a long time, because I have wasted about a year messing about with this sloppy setup. There are so many many things wrong, and there is no way this paper will be retracted.