I have just read Dave Lunt’s text on Anonymous peer review, and I can only second every word he writes. I consider reviewer anonymity as essential for giving me the freedom of saying things that have to be said. Like David, I would never dare telling a person I know very well that their latest manuscript is essentially rubbish. Without any doubt, my case history as a reviewer would have severly limited my chances of later collaboration.  Even very good scientists can sometimes submit very bad manuscripts.

The other side of the coin, giving ‘rogue reviewers’ the chance to hide behind anonymity has to be addressed, although I would rather see this happen without compromising anonymity. I even have an idea how this could be done. I am sure that some clever people have proposed this before. I wouldn’t even exclude that some (other) clever people have given the idea due consideration and found out that it is not going to work. Anyway, this is my suggestion: Journals publishing peer-reviewed research should have a formalized ‘court of appeal’. The authors should have the possibility to do something about reviewer reports that are way off the mark, be it due to hostility, incompetence, or laziness of the reviewer. Obviously, safeguards have to be established that prevent authors from complaining about every negative review they get. However, authors with justified complaints must be guaranteed that some knowledgeable and unbiased person sees to their case. In an ideal situation, the board of appeal could not only overturn the original judgement, but also do something about egregious cases of flawed reviews - something that would actually hurt the reviewer (no, I am not thinking of high voltage here).

Already now, it is possible to complain with the editor about reviewer’s misjudgements, at least at some journals (generally the better ones). However, unless you know the editor in person, or are a big shot working in the right place, your chances of getting the editor to even look at your case are slim. Given that nobody likes getting a bad review and is likely to find the criticism unjustified, the lack of enthusiasm on the editors side to deal with complaints is understandable. However, there are really some bad bad reviewers out there, and something has to be done about this problem. Nowadays, when more and more journals are charging the authors rather than the reader, the paying ‘customer’ should at least have the right to be treated in a fair and transparent manner.

Probably, the two weakest points in my proposal are i) the increased work load put upon the editor(ial board), and ii) the lack of adequate sanctions that can be imposed on bad reviewers. After all, we have to keep in mind that reviewers are not paid for their job, and every action that risks scaring away the good guys is out of question. Maybe, in the more severe and unambiguous cases, at least a stern letter from the editor to the reviewer, mentioning the flaws uncovered during the appeal process, would be indicated. This action would not destroy the career of a bad reviewer, but might nudge him/her back on track.

Yesterday, I have received two ads - disguised as blog-comments - from Walter Jones, apparently the AAAS expert for viral marketing. While one copy would have probably done the job as well, the topic was interesting and pertinent to my blog. Thus, I am willing to assist the AAAS and make their announcement known to a wider audience: my blog readers. Both of them.

In their Webinar series, Science Magazine will offer on May 1, 2008 a webinar entitled “The Ubiquitin-Proteasome Pathway: Targets for Disease Treatment and New Tools for Discovery“.

This topic is very similar to that of a conference I have attended recently. There doesn’t seem to be an overlap of speakers, though. As the announcement says, there will be three speakers: Alfred Goldberg from Harvard Medical School (who had also been announced for the Berlin meeting but did not show up), Barry Schweitzer from Invitrogen (the company sponsors the webinar and apparently hopes to sell more of their protein arrays), and Paul Andrews from Amgen. The whole thing will be moderated by Sean Sanders from the AAAS.

Tackling the UPS for drug discovery is an interesting, but also rather difficult area, and I am not yet convinced that will see any UPS-targeted drug within the next decades. Anyway, the topic is close to what I am working on, and I wil probably attend…. if it works for me. The technical provider of the webcast (event.on24.com) offers a page to test your system for webcast capabilities, and mine failed on several counts. For once, the report says that my operating system “fails”. They do not appear to like linux but recommend “Windows XP SP2 or Mac OS X 10.4″ instead. In addition, our company’s leased line “does not meet the Minimum Bandwidth Requirement for viewing this content”, hmm.

Anyway, I will try getting it to work anyway.

Here are just two interesting stories I read on other peoples blogs:

First, Jake Young at Pure Pedantry blogs about a recent Cell paper by Sakaue-Sawano et al. who present a clever application of protein ubiquitination for visualizing the cell cycle stage of cells in vivo. Lars Juhl Jensen at Buried Treasure has also picked up this story, and those two blogs provide a lot of detail on the method, including a link to a nice video showing HeLa cells passing through 3 cell cycles. In brief, the authors of this paper exploit the fact that several protein ubiquitination systems are only active during particular phases of the cell cycle. On one hand, there is the APC/Cyclosome system, which degrades target proteins only in late mitosis (APC means Anaphase Promoting Complex) and in G1 phase. Conversely, the SCF-Skp2 system is mainly active in S and G2 phases. By coupling two different dyes to target proteins of APC/C and SCF-Skp2 respectively, it was possible to observe an oscillation between red and green colors as the cells go through the different cell cycle phases.

On a very different note, Peter Murray-Rust blogs about the lack of data mining possibility in the Pubmed Central resource. He goes on to discuss if - in the light of this shortcoming - PMC can still be considered an open access resource. I can understand his concern and would certainly welcome if PMC and other open access scientific repositories can be used for automatic text mining efforts. However, what really struck me was the statement:

When George Bush signed the mandate he clearly envisaged that the information should be used for the benefit of human health…
…and this means text-mining.

I am not sure how serious Peter was when writing this. I hope not too much. I cannot think of a single example where text-mining has ever made a major contribution to solving any real-life biomedical problem. Even if there are such eamples, their number will be small. If we compare the health benefits from text mining efforts to those provided by real (human) scientist reading the literature, I have no doubt that the latter would prevail by a big margin.

There should be no doubt about it, it would clearly be a good thing to enable text-mining on PMC. However, describing the current situation of free access to PMC papers for scientists as useless without added text-mining capabilities appears to be, well, kind of biased.

Reading an interesting post on the tree of life reminded me that today is April 1st. I will seize this opportunity to make you familiar with an interesting protein that had been discovered in - you might have guessed it - April, exactly 10 years ago. In reality, the name APRIL has nothing to do with the month but rather means “A PRoliferation Inducing Ligand”. You can look it up somewhere in JEM.

APRIL is a member of the TNF family, which also contains proteins such as TRAIL and Fas-ligand. Unlike many other TNF family proteins, APRIL does not induce apoptosis, quite to the contrary: By a mechanism that is still very incompletey understood, APRIL upregulates the anti-apoptotic members of the Bcl2 family, most likely by way of NF-kB. Initially, we thought that APRIL had a connection to tumor-specific proliferation events. We had seen APRIL message and protein predominantly in tumors and only very little in normal tissues. Moreover, when various tumor cells were treated with APRIL, their proliferation rate was enhanced. I haven”t followed this field very closely, but I am under the impression that this tumor connection is shaky at best. A quickPubmed scan even found several reports of APRIL being stronger expressed in normal tissue as compared to a matched tumor.

Nowadays, it seems that APRILs role is mainly in B-cell biology, similar to its younger (but better-known) brother BAFF (some people call this protein BLyS, but I won”t). One of the more intriguing features of BAFF is its name. My initial choice was BPRIL (leaving room for CPRIL, DPRIL and so on), obstensibly an acronym for “B-cell proliferation inducing ligand”. As usually, nobody cared for my name choice, and in the end we “agreed” on BAFF (for B-cell Activating Factor - can”t remember where we got the 2nd F from.). At some point in time, when a patent application on this protein had to be filed, we had a name vacuum, resulting in an obscure application with the even more obscure title “Kay - A NOVEL IMMUNE SYSTEM PROTEIN”. As you would expect, this application was rapidly superseded by something more decent, but it can still be accessed at freepatentsonline.com. I just love the claim on “modified Kay-ligands and pharmaceutical compositions comprising them”. In the meantime, there is no shortage for names to choose from: BAFF is also known as BLyS, CD257, TALL, TALL-1, THANK, UNQ401, ZTNF4, TNFSF20 and TNFSF13B.

So, what does BAFF, the B-cell activating factor, do? Right. APRIL seems to work similarly, which is maybe not too surprising, given that the two ligands are relatively similar and share the two receptors BCMA and TACI. A third receptor, BAFF-R, seems to be specific for BAFF, while BCMA has been reported to prefer APRIL over BAFF. Nevertheless, BAFF/BlyS is better known, as it is the target of Belimumab (Lymphostat-B), a monoclonal antibody which is being developed by HGS as a therapy for rheumatoid arthritis (RA) and possibly other auto-immune diseases. As far as I know, APRIL has not specifically been targeted for drug development, although there are early clinical studies of Atacicept, a TACI-Ig fusion that would be expected to inhibit both APRIL and BAFF.

What is even worse, the drug people tend to use the ‘wrong’ name for BAFF/BlyS: Instead of writing reviews with titles like “A BLySful end to autoimmune disease”, people should think about “A BAFFling success in autoimmune therapy”.

Before I forget: To properly celebrate April 1, I have included one factual error in the above text. Spot it and earn an (BlySful, BAFFling, extra-TALL) beer at the next conference. At the moment, this post does not include any links, they will be added later.

This is today’s 2nd post on recent papers describing the beneficial effect of things gone wrong. The first one was from the area of bioinformatics, this one is about gene silencing by RNA interference (RNAi). A lot is known about the mechanism and application of RNAi, see e.g. the Wikipedia entry. Basically, you can use RNAi to downregulate a gene of interest by designing a short oligonucleotide (siRNA) complementary to the target mRNA and getting it into the cell to do its magic. Inside the cell, the oligonucleotide pairs with the target mRNA and the resulting duplex RNA is recognized by a specific cleavage machinery, which is present in the cell anyway. Most likely, this RNAi cleavage machinery has not been put into the cell by an intelligent designer to facilitate our lab work, but rather serves as an antiviral defense mechanism: double-stranded RNA is not normally observed within the cell and is interpreted as a sign of viral infection and thus removed.

It is somewhat naive to assume a clean downregulation of just a single targeted gene. One of the things I am doing to pay my rent is the bioinformatical analysis of microarray data. In this context, I often encounter transcriptional profiles that are the result of RNAi experiments. What you typically get is a mixture of the following effects:

• Target effect. This is what you want to see - your gene of interest is downregulated.
• Off-target effects. RNAi does not require perfect matching; your siRNA oligo will probably form imperfect duplex strands with several other (unwanted)  mRNAs leading to their downregulation
• Secondary effects. Downregulation of the target genes (both wanted and unwanted) might lead to expression changes in other genes. Particularly pronounced when your target is a transcription factor.
• Unspecific dsRNA response. Cleavage is not the only antiviral response to double-stranded RNA. Many dsRNAs also stimulate the innate immunity pathway, e.g. mediated by toll-like receptors (TLRs).

The paper I am talking about has just appeared on Nature’s AOP page. It is from Mark Kleinman et al, the title is “Sequence- and target-independent angiogenesis suppression by siRNA via TLR3“. It is a very instructive story and goes to show that controls are a very important thing to do. The original idea was to to design a 21bp siRNA targeted against the vascular endothelial growth factor VEGFA with the idea to prevent a neovascularization in the eye that can lead to blinding.  Apparently, the RNAi experiments did show the expected effect. However - surprise surprise - it turned out that this effect was completely unrelated to VEGFA downregulation and could be mimicked by any old siRNA. It looks like the last confounding factor in my list was the dominant one: in the studied system, dsRNA lead to an innate immune response mediated by TLR3 , with the typical  concomitant  induction of gamma-interferon and interleukin-12.This just emphasizes the notion that getting the desired effect does not necessarily mean that your experiment works. Or, as a former labmate of mine would put it, “if your experiment gives the the expected result first time, there must be something seriously wrong”.

Note added in proof: I just saw that Eric at the Futile Cycle has also reported on the same story.

This week, I found two strange but interesting papers, one from the area of bioinformatics, the other from gene silencing. The unifying topic is that in both published works something went wrong, but the results were nevertheless as good - or even better - than intended.

The first paper, published in the latest issue of Nature Biotechnology, is from Mark Styczynski et al. The title is “BLOSUM62 miscalculations improve search performance“. I must say that I was very surprised to see a paper like this. BLOSUM substitution matrices are widely used throughout bioinformatics. In particular the BLOSUM62 version is popular as it is used per default in NCBI’s protein BLAST applications. I am a heavy user of BLOSUM matrices myself; on my computer it is the BLOSUM45 matrix doing most of the protein comparison work. In case you wonder what I am talking about: When you want to compare protein sequences, it is not good enough to just score identical or non-identical residues - you need some measure of amino acid similarity. For this purpose, you (or rather your program) turns to a ’similarity matrix’, a.k.a. ’substitution matrix, which tabulates a similarity value for each possible amino acid comparison. Initially, a number of different similarity matrices have been tried. It soon turned out that the best performance was obtained from matrices that do not score some kind of physical similarity between amino acids, but rather the frequency of one amino acid being substituted by evolution for another one. For example, you often find Leucine replacing Isoleucine (resulting in a good score for the pair Leu:Ile) but you rarely find Leucine replacing Aspartate (resulting in a bad score for Leu:Asp). In 1992, Steven and Jorja Henikoff analysed a large number of amino acid substitutions in a database of sequence alignments at varying evolutionary distances. The result of this analysis was the BLOSUM series of substitution matrices, with the most popular BLOSUM62 being derived from sequence alignments with more than 62% sequence identity (if I remember correctly).
Now, 16 years later, Styczynski et al apparently found a bug in the program used by the Henikoffs to calculate the BLOSUM matrices. In their paper, the authors describe a re-calculated ‘correct’ version of the BLOSUM62 matrix and compare it with the ‘wrong’ Henikoff version for its performance in sequence comparison applications. Most surprisingly, they report that the original  BLOSUM62 version performs better than the correct one!
Here is a quote from the Styczynski paper:

We find it interesting that the BLOSUM62 matrix is used every day (and more interesting still that its derivation is a common topic in computational biology classes), and yet we can find no previously published mention of any of the errors discussed here. We did find that some of the errors were fixed in later tangential work by the original authors, but the ‘correct’ matrices have never been published or adopted. We also note that the existence of statistically significant improvements due to (essentially random) software errors supports the notion that there is significant room for improvement in our understanding of protein evolution. Of course, software errors are quite common and nothing special; however, it is at least a curiosity that these errors stayed buried for so long and have been improving BLAST searches (ever so marginally) for the past 15 years.

Maybe someone should have a look at the BLAST sources and try to find the bug that makes BLAST perform so well.

It looks like the guys at SART take longer than expected to finish their groundbreaking research. This leaves me some time to blog about another topic that has caused some stir in the blogosphere recently. I just was reminded of the story when reading an editorial by Donald Kennedy published in last weeks Science.

What seems to have happened is that Pfizer is being sued on side effects of Celebrex and Bextra, and some material published in NEJM is to be used as evidence in this case. It now seems that Pfizer tries to force NEJM by means of subpoenas to let the Pfizer lawyers browse confidential NEJM review material. I am neither familiar with this particular case, nor with the relevant US laws. Here is a short paragraph from the editorial:

Pfizer asserts that in some cases plaintiffs are making use of published papers from the New England Journal of Medicine (NEJM). So it wants to dig though the confidential reviews of those papers in search of something to strengthen its defense.

You can find extensive coverage of this incident in a number of blogs, including Pharmalot,  The Futile Cycle and at Nature network. As can be expected, bloggers and most scientists I spoke to take side with the NEJM. As a (moderately) frequent author of scientific papers, and a (more) frequent reviewer, I also feel inclined to fiercely defend reviewers confidentiality - at least against lawyers and other low life forms.

However, a closer look at the wording made me hesitate. If Kennedy’s editorial is correct, NEJM is not asked for the reviewers identity, but rather for ‘confidential review material‘. And this is - at least in my humble opinion - an entirely different matter. One feature of peer review forms at most journals is a box with the title ‘confidential remarks - not to be sent to the authors’, or similar. I never quite understood what these confidential remarks are good for. As far as I can remember, I never used this feature myself. In general, I think that peer review should be made as transparent as possible. I would strongly prefer that decision about life or death of a submission should be made on the basis of material that can be seen (and understood) by all parties involved. The prospect of some dark collusions going on between editors and reviewers is nothing that any author would fancy.

I have been raised in the scientific working class, i.e. a minor university in some obscure European city. In these circles, you will often find a deeply rooted distrust in the editorial processes at high-profile journals. I am not saying that this distrust is generally justified, but it is rather widespread. Many people won’t even consider sending their good stuff to Nature, Science, Cell. The general feeling is that these journals are exclusively meant for PIs working at prestigious places, who are used to having dinner with chief editors, which whom they are on a first-name-basis anyway. Ah yes, and a flawless scientific pedigree is also required. In addition, everybody knows somebody who can tell a sorry tale of what happens when you dare to swim with the big fish.

To make this posting more interesting, here are my own two contributions to the elegy of the scientific underdog. I must admit that I wasn’t directly involved, but I can guarantee that these are not just urban legends. In one case, I have seen black-on-white evidence, and the other case happened to somebody I consider very trustworthy (and I was a minor co-author of the paper).

Case 1: This incident happened about 15-20 years ago, but trust me - both the scientists and the journals involved are still around. A colleague of mine, I will call him X, submitted a paper to a moderately high-profile journal. It is important to know that at that time manuscript submissions were not done via the internet, but sent by regular mail on real paper. The journal in question offered the option to exclude particular competitors from the reviewing process. I don’t even know if journals are still doing this, but I consider it a nice feature (although i never used it myself). X had a particular competitor Y , who was working at a far more prestigious institute than X himself, and X asked the editor to exclude Y from the reviewing. The big surprise came when X received the returned manuscript (yes, at that time you got the manuscript back, mainly because of the glossy-print figures that had to be supplied by the authors). I don’t remember if the paper had been accepted or rejected, but what I do remember is finding - hidden between the pages of the manuscript - a signed letter from Y addressed to the editor, saying ‘Dear xxxx, thank you for bringing this manuscript to my attention’. I also remember that I thought this was outrageous and never understood why X never did anything about it.

Case 2: This is not quite as bad as the previous case, but it happened recently. Another researcher ‘X’ (different person, obviously) submitted a manuscript to a high-impact journal Y. It did go out to the referees but eventually was rejected on the basis of two mostly negative reviewer reports. This was not entirely unexpected, and in the meantime the manuscript has been published by another decent journal. To X’s surprise, at a recent meeting X was approached by another scientist Z, who asked why the manuscript hadn’t been published in the original journal Y. During the following discussion, Z admitted to being a reviewer of the original submission to journal Y (we all know that Z is not supposed to mention this, but we are all humans, even reviewers). When asked which of the reviewer reports was his, Z answered ‘the enthusiastic one’. Strangely enough, both of the reports were far from being enthusiastic, so chances are that Z didn’t tell the truth. However, X had wondered all the time why the two negative reports used for the rejection were titled ‘reviewer 1′ and ‘reviewer 3′. Did the journal really drop the (possibly positive ) report of reviewer 2? And if so, did they do it on purpose? Probably, they even are entitled to do whatever they want. Most journals have policies saying that rejection or acceptance is solely a matter of the editor’s discretion. In any case, I consider it bad style to make editorial decisions on the basis of material that is hidden from the authors.

I should hasten to add that I also had a good deal of positive experience with the reviewing process, both as an author and as a reviewer. By no means do I consider the two cases typical of what is going on at the major journals of the field. However, both cases serve to underscore my point that the obscurity and secrecy of the editorial decision process should be kept at a minimum. But not below that.

Seems I have neglected this blog for too long and you, my dear audience, have turned to greener pastures. Well, this calls for some drastic action. You have been warned

Just a brief announcement: On March 11-12 2008, there will be a conference on drug targets in the UPS taking place in Berlin, Germany. Judging by the topic and the organizer (Informa) it is safe to assume that the conference will mainly cater for scientists in pharma and biotech companies. As you know from my previous post, industry people are swimming in money, and thus the conference fee is a bit on the expensive side.

As there isn’t much UPS industry in Europe, the invited speakers are mainly from the US. Amongst others, there will be presentations from Genentech (Peter Jackson), Millennium (Joseph Bolen), Roche (Lyubomir Vassilev), Novartis (Martin Renatus), Merck (Nathan Bays), Bayer Schering (Bernard Haendler), Progenra (Ben Nicholson), LifeSensors (Tauseef Butt), Hybrigenics (Laurent Daviet) and Nereus (Michael Palladino), so there are some interesting talks to look forward to. Ah yes, and they invited yours truly, supposedly as some kind of counterweight. If I read the program correctly, my talk will take place during dinner, which will probably help the audience to survive 40min of ubiquitin bioinformatics. There will also be a few presentations from academia, amongst others Alfred Goldberg (Harvard), Huib Ovaa (NKI Amsterdam) and Martin Scheffner (Konstanz).

Just in case you wonder: No, I don’t get paid for advertising this conference. It is just that - working in the biotech industry myself - I like the idea to have a conference on the drug target aspects of the UPS, especially here in Europe. It is a pity that the meeting is that expensive, but this probably explains why the conference is still not fully booked.

I really shouldn’t write a blog posting right now, but rather work on my slides for an upcoming conference. Alas, working on slides is not my favorite occupation, and now I find myself pondering questions like “Is inviting me for talking at a conference a service for me or for the audience?”. A strange question to ask, maybe (unless you have heard one of my previous talks and thus can safely discard the idea that it might be to the benefit of the audience). Here is what made me think about this problem in the first place:

On previous occasions (e.g. in my blog, but also in the discussion of other people’s blogs) I have complained about software and web servers that are free to use for academics, but which charge an (often horrendous) amount of money if you happen to work for a company. For obvious reasons, I don’t like this practice, but I can understand the underlying logic: The software authors or right owners assume that my usage of their service helps me to earn big money, so it is just fair if they charge me big money as well. Today, I noticed a similar arrangement for the use of copyrighted material at academic conferences. In this particular case, however, I fail to see the point.

An hour ago, I browsed the contents of the new issue of Nature (online, of course), more specifically this article by a group I used to work with in the past. At that point, I noticed for the first time a small link button entitled ‘rights and permissions’. Probably, these links have been around for quite some time, but so far I must have missed them. With the nagging thought of imminent slide preparation in the back of my head, I got curious and thought “well, let’s assume I want to use a figure from this paper in my talk. This link will probably tell me what I have to do for getting a permission”. I clicked the link and answered a few innocuous questions like “what to use? figure! “, “how many? one!”, “what purpose? academic conference!”. When I arrived at the question “requestor type?” there was this sinking feeling that this story would not come to a happy ending. My first impulse was to lie (it is just a web form, after all), so I checked “university/public research institute”. After submitting the form, I got the encouraging answer

This reuse request is free of charge although you are required to obtain a license through Rightslink and comply with the license terms and conditions. You will not be charged for this order.

Ok, that is not so bad. I have seen many people using figures from published papers in their talks. I have no idea if they really asked for permission, but if they did - at least they didn’t have to pay for it. My next experiment was to go back to the form, this time staying closer to the truth and answering “Other for-profit company” as the requestor type. Upon submission, I learned that I would have to pay $690 for showing the figure in my presentation. Or $747.50 if I want to use the high-res version. I briefly checked if there was a discount if I requested more than one figure, but no, there wasn’t. At least, NPG is candid about their conditions and doesn’t try to evade by saying ‘contact our sales department for a quote’.

Sorry Jürg, no figure from your paper will be shown in my talk. I just cannot afford it. Didn’t fit the topic anyway. But now, instead of eagerly starting to draw some figures, I find myself wondering why showing a measly blot or schematic must be that darn expensive? Le me start with a simpler question: why would a publisher want to control the amount of copyrighted figures shown at an academic conference? The standard reasoning, I guess, would go: There are lots of people in the room who get to see the figure and thus do not have to buy a copy of Nature anymore. Or something along these lines.

But then, why does it make a difference if the talk is presented by somebody working at a university or at a biotech company? The only logical reason for this distinction would be if the beneficiary of showing the figure in my talk would be me rather than the audience. I am not sure how publishers would argue here, maybe ‘by presenting this copyrighted figure in his conference talk, the speaker will be enabled to earn lots of money (travel reimbursement), so we deserve our share of it’. Or, alternatively, ‘these biotech guys are made of money, so why not ask for $690, it is just peanuts for them’. Either way, with this price tag, I have no choice but to draw all figures myself.

P.S. I have just checked JBC, which uses a similar form for permission requests, but arrives at a more moderate price of $33 for an estimated audience of 250. Unfortunately, limiting the number of conference attendants to 12 (closer to my average audience) doesn’t help to cut the cost any further.