Achieving Reproducibility: Don't Let Antibodies Be a Variable in Your Experiment

– [Robert] Hello everyone Welcome to today’s live broadcast, Achieving Reproducibility: Don’t let Antibodies be a Variable in Your Experiment I’m Robert Castilanis of LabRoots, and I’ll be your moderator for today’s event We are delighted to bring you this educational web seminar presented by LabRoots, and sponsored by Cell Signaling Technology Cell Signaling Technology, CST, is a private, family-owned company, founded by scientists and dedicated to providing the world’s highest quality, innovative research, and diagnostic products to accelerating biological understanding and enable personalized medicine Our employees operate worldwide from our US headquarters in Massachusetts, and our offices in the Netherlands, China and Japan For more information, please visit CellSignal.com We have a few important announcements before we begin The webcast is designed to be interaactive, and we encourage you to ask questions during the event You can submit questions by typing them in the Q and A box, which can be found by clicking on the green Q and A button at the lower left of the presentation window We’ll try to answer as many questions as we can You can enlarge the slide window by clicking on the screen icon in the lower right-hand corner of the slide window If you have any technical problems hearing this presentation please click on the support button at the top right of your presentation window, or submit your problem through the green Q and A button, lower left This is an educational webinar, and this offers three continuing education credits After the webinar is over, please click on the CE button located at the bottom left-hand corner of your web page, and follow the process of obtaining credits I would like to introduce today’s speaker, Dr. Anthony Couvillon Dr. Couvillon has worked at Cell Signaling Technology for almost seven years in various capacities As a product development scientist, he conceived, validated, and released over 100 antibody products More recently, Dr. Couvillon was in charge of the CST’s unique post-transitional modification and motif antibody portfolio, and has considerable experience in the proteomic space Prior to joining CTS, Dr Couvillon earned his PhD in molecular physiology and biophysics at Vanderbilt University, and spent many years as a postdoctoral fellow at the Harvard Medical School He had dedicated much of his career using enveloped infinity-based methods (mumbling) Currently, Couvillon is responsibLe for coordinating CST’s antibody reproducibility efforts, working with both internal and external scientists to ensure the CST meets or exceeds the industry standards for an antibody validation I’ll now turn it over to the doctor for his presentation – [Anthony] Hi everybody, and welcome to the webinar Again, my name is Anthony Couvillon, and what I’d like to talk to you guys about today is how to use antibodies, and what CST’s take is on antibody validation and reproducibility The way I’d like to do that is first talk a little bit about CST, about myself, and why you should even believe what I’m telling you today But then go back a little bit, and talk about the reasons for why reproducibility is an important topic, what the causes are, and who’s responsible The CST’s role in validation and reproducibility, and then discuss a little bit about what’s being done to address the problem, not only from CST’s perspective, but also from the NIH and other players At the end, we’ll take some questions As we just heard, CST is a private global company It was founded by a group of PhDs from MGH and Harvard, with the goal in mind of generating reagents, useful reagents that researchers could use to enable pre-clinical discovery CST is focused on application-driven antibody validation We put a lot of effort into quality control and our tech support Internally, we’ve developed three generations of rabbit monoclonal technologies, and that leaves us to be able to generate recombinant rabbit monoclonal antibodies directly from polyclonals, and that’s our latest technology As a matter of fact, 95% of our rabbit monoclonal antibodies are indeed recombinant We do all of our manufacturing in-house That is 100% of the antibodies that we sell are manufactured right here in Danvers, Massachusetts

in an ISO9001 certified facility We develop over 95% of our products Just briefly, a little bit about me, just so you know where I’m coming from in this area, I’ve been working in labs for a very, very long time First as a research associate, then as a graduate student and a postdoctoral fellow, like many of you I’m sure have Now the interesting thing about this is that every time I’ve worked in one of these fields, I’ve been in very antibody-dependent labs Everything from studying serine/threonine kinases at Harvard to G-Proteins and GPCRs at Vanderbilt during my graduate career The thing that always struck me in each of these studies, as I’m sure many of you are familiar with, is that oftentimes, antibodies are one of the limiting steps in your hypothesis-driven research Either there’s issues with lack of availability of a specific antibody, the specificity of the antibody that you buy from a vendor, or it’s not clear exactly if the antibody works for the application you’re interested in Or you transition from being a graduate student to a postdoctoral fellow, and you change methodologies You’re using techniques that you’re no longer familiar with In my case, I went from never doing flow cytometry, to doing almost exclusively flow cytometry There was a learning curve in there in understanding how to use antibodies that had been used widely by others, that I just couldn’t get to work in my own hands The bottom line was that I spent two years of my life in graduate school using an antibody from a company that turned out to be non-specific Part of the fault was mine for not taking the time upfront to validate the antibody to make sure that it was specific upfront Oftentimes, as many of you know, when you’re in a lab and your boss is asking you to churn out data for publication or whatnot, you trade off production over taking that second step and wanting to make sure that the reagent is correct Part of the reason that I came to CST, and one of the reasons that many folks are here, is because we’ve all had that experience We’ve all been in a lab where we worked with an antibody that has turned out to be incorrect, and you end up wasting time, resource, and your own sanity, in order to get your experiments to work In coming to CST, like I said, I’ve been here for seven years Much of that time has been spent doing antibody development That’s everything from selecting the targets, to assisting with the design of the antigen, to doing all the validation work that goes into producing a product that then is bought by our customers I’m showing here just a few examples of the antibodies that I developed I’ve developed probably over 100 antibodies, mostly in the mAb kinase, calcium signaling space, but a few outliers in there, as well In every case, I didn’t want a student, or a user of my products, to have to go through the issues that I went through Everything that I did, and everything that we do here at CST is all about ensuring that we make the highest quality product possible We want people to be able to use our antibodies right out of the box, as they would expect them to work Part of the problem with reproducibility is the way that we choose antibodies The way that I see it is that there are a few different options when you’re first looking for an antibody You can go to the scientific literature and try to see how an antibody was used You and I all know that there are two issues you face The first is that you find the manuscript using Google Scholar or PubMed, and you locate the antibody, the experiment you want to use, and the authors don’t actually cite it My favorite case of course, is the one where they say, oh it was used exactly it was used in the previous paper You hunt the previous paper down, they then reference another paper, and so on and so forth One of the problems with the scientific literature is that there are no citations, the methods are incomplete, or there’s no details about the antibody that was used The other place you can go is to look at producers websites As many of you have probably known, producers either don’t offer the product, or it’s very difficult to understand how they validated the antibody Companies can show data, but data showing staining

in a cell is not validation That’s just showing you that the antibody detects something Doesn’t tell you exactly what it is Some companies do a better job than others of showing that kind of validation data At CST, we hope that we’re extremely transparent We’re not perfect, but as we go, we do a better and better job of showing exactly the data, exactly the systems that were used to validate the antibody, and I’ll discus that in a lot more detail in the future One thing to be cautious of, and something that’s kind of popped up more and more lately, are the presence of resellers If you often go to Biocompare, or any other antibody search engine, a lot of times the places that pop up are companies that are just reselling other companies’ products They don’t ever make it clear unless the clone name is the same that the product is exactly the same as what the other company is offering We’ve heard from many, many different customers that when they set out to do a new product project, if they want to stain a new target by IHC for example, they’ll go out and they’ll buy five or six different antibodies from five or six different companies, and they’ll test them out The reality is that depending on the target, three or four of those antibodies could be exactly the same product formulated slightly differently, but basically you’re being resold the same product in a rebranded way There’s a lot of caution that needs to be used when purchasing antibodies from what I will call a non-producer, or a reseller There are also search engines and aggregators that you can use to look up antibody information Antibody Resource, Google, Antibodypedia for example, and there are a lot more and more of these that are coming out, and I think that these are beginning to be a really useful resource for people to find antibodies Biocompare is probably one of the most used A lot of times, what you have to realize with these sites is that the vendors are paying for placement of some of this information on their sites The top results that turn up are not always the best antibodies, they’re just the first ones that were listed A lot of these sites do have reviews, which I think are helpful, but you have to use caution when doing that Did the review show data, for example? The other thing is, and we’ve seen this as well, is that some of these sites will often aggregate the data from a bunch of different vendors so that you can never quite tell which data is associated with which antibody I think that that’s confusing, and creates a problem You can also often reach out to your colleagues and peers, and similar to digging into the scientific literature, the problem that you have is that people may not remember exactly which antibody they used More importantly, they may not remember which lot they used, and there are variations, especially with polyclonal antibodies, from lot to lot You have to be really careful of that Why is this important? Some of you may be familiar with the Global Biological Standards Institute, which is headed Len Freedman Dr. Freedman did a really interesting study with a group of economists a few years back in which they looked at scientific reproducibility in the preclinical environment and made some estimates, both bottom up and top down, to try to understand how much of academic research is not reproducible They estimated, based on a number of studies, that over 50% of all preclinical research is not reproducible That amounts to a waste of $28.2 billion in NIH funding that’s given to academic researchers every year Now that 50% number may vary Other studies have shown it to be higher Other studies have shown it to be lower In general, 50% is a pretty good estimate of how irreproducible research is It’s not clear how much of that is due to antibodies, but for the most part, what Len Freedman’s group found was that biological reagents and materials were probably the largest fraction of waste That is everything from identification of the reagent, incorrect reagents, bad reagents, poorly validated reagents, reagents used in the wrong system, and they found that that made up the bulk of the waste, bulk of the cause for irreproducible research They also found that investigators who didn’t set their studies up properly, or cherry-picked data, or didn’t do enough biological replicates also contributed pretty highly to that number Study design, and the data analysis, and report out

were also a large chunk of that irreproducible research Finally, the use of protocols That is that protocols that are not completely written, or don’t document everything, or are just in somebody’s head were the smallest cause, but still a significant cause of that lack of reproducibility This is important because the research that’s done in the preclinical environment is what drives the discoveries that go on in the clinical environment If the information’s being developed in a preclinical phase of a research project is mistaken, or incorrect, or incomplete, then it means that we’re slower getting to cures, we’re slower getting to discoveries, or worse yet, we’re actually making bad diagnostics, bad therapeutics, based on incorrect data There has been a history of this happening A couple of recent studies, again one from the GBSI, and another one from Monya Baker’s group at Nature, did a series of surveys Out of those surveys of scientists, research scientists, they found that there were a number of causes, general causes that contributed to irreproducibility within the lab I’ve listed them in order of importance here The survey results indicated that incomplete reporting of data was the number one That means cherry-picking information Reagent variability and lack of validation were number two and number three Reagent variability is lot to lot consistency, reseller’s consistency, people using different versions of the same thing, or not knowing exactly what they’re using, or not doing validation That’s a really big problem if those two things are right up there at the top Experimental bias also came up People not understanding how to do enough biological replicates, doing experiments without statistical basis, or selectively reporting the results, made up a large chunk of the issue, as well Then again, there was incorrect interpretation and misconduct were by far the smallest factors in this What it really comes down to, at least for me in cell signaling, is that we can break all of these down into three basic categories The way we feel, and a couple recent meetings would support this, and I’ll talk about those in a bit, is that really all of the issues come down to three problems Inadequate training and support, little or no reagent identification or methods reporting, and poorly validated or misused antibodies At CST, we call those the three M’s of reproducibility I’m gonna walk through this very, very quickly We feel that reproducibility comes down to three things, methods, materials, and mentoring I’ll talk about mentoring first CST feels that one of the biggest problems is that there’s a lack of basic training and a lack of basic knowledge of antibodies We see this with our own tech support, which is that we have users call in that are just unfamiliar with how to use an antibody Especially younger investigators, this was also reflected in the GBSI report, aren’t familiar with how to validate an antibody, or don’t see the need to validate an antibody While the vendors should do a lot of that work for you, the bottom line is that you still need to validate the antibody in your own system in your own hands to make sure that it’s gonna work in your specific system I’ll talk more about that in a bit There seems to be a broad misunderstanding that an antibody that works by Western blot doesn’t necessarily also work for IHC, or another application If there’s one message you take home from this webinar today it’s that performance in one application does not predict performance in another Whenever you take an antibody and use it in a different application, you need to revalidate that antibody I’ll talk about that in a lot more detail in a little bit Again, students and scientists learning how to understand and interpret their data, set their experiments up correctly, and using the same methods over and over again, and understanding the difference between biological replicates and technical replicates, was cited in both studies as being a problem Finally, we feel that tech support is a component of mentoring, and that when you have an issue with a company’s reagent, whether it be an antibody or anything else,

you should be able to call that company and get help Whether it be guidance, a protocol, whatnot, that companies should play a role in that mentoring process Methods on the other hand, fall mostly unto making sure that they’re clear and detailed They’re complete, but they’re available, traceable, and that they can be replicated or repeated This is really important, and I’ll talk about this when I talk about the role of journals, scientists, and vendors in this space in a slide Finally, the materials that you use to do your experiments need to be highly validated You need to use them properly, as I just alluded to, that an antibody performance in one application doesn’t predict how it’ll perform in another The materials need to be easily identified They need to be consistent, and that could be lot to lot consistency, or that could be use to use consistency, and they need to be stable I’ll talk about those concepts in a second Who’s responsible for each of these areas? What I would argue is that principal investigators and their institutions should be responsible for the mentoring and methods component of the three M’s That is that principal investigators and institutions should be training their students in statistical analysis, and how to set up experiments, and the proper use of reagents and protocols They need to be able to document those to show that, they need to just teach basic good laboratory practices, and keeping notebooks, and things like that If principal investigators and institutions play a greater role in here, there would be less of a reproducibility issue We realize that this is difficult for a lot of investigators to do They’re busy writing grants, et cetera Still, we feel that this is an important initiative, and I think that coming into the future, there will be more and more effort by institutions to help train, especially younger scientists Scientists, or the people using the reagents, are responsible for the materials and methods You could also argue they are somewhat responsible for mentoring those underneath them For this purpose, I am going to make them predominantly responsible for the materials and methods The materials and methods means that you should again, identify the materials that you’re using, clearly document the protocols that you’re using, and make sure that other people can replicate the experiments that you’re doing I think that that’s very important The scientists are a little bit more careful with how they use reagents, and how they document how they use reagents Some of this reproducibility situation will be mitigated We also think that journals are complicit in ensuring that materials and methods are clearly documented and clearly identified in their journals As I alluded to early on in the talk, there are all sorts of times in which you go to a journal, there are clearly lots of Western blots in the article There are clearly lots of Western blots in the article Sorry, I think we’ve gone blank There are clearly lots of Western blots in the article, and the Sorry, forgive me There are lots of Western blots in the article, but nowhere are there citations indicated for the antibody It’s important that the journals establish rules and guidelines to show what you need to do to indicate, properly cite, and properly reference an antibody when you use it At CST, we feel that CST itself and vendors are responsible for all three of the M’s That is that we need to make sure that we carefully document our methods, we need to make sure that we clearly identify and make our, the materials, the antibodies that we’re providing you, extremely consistent, and that we’re providing you tech support so that we’re mentoring and assisting with the experiment How do we do that? At CST, we use application-driven antibody development, which means that all of the protocols that we show on our website have been optimized for that application for that antibody We make those detailed protocols clearly available on our website in an app-specific manner So that if you go to our website and you’re using a specific product, you can find out more information We have significant data to show that even changing the percent of formaldehyde in the fixation conditions of an IHC or an IF experiment can have disastrous effects on the outcome of the experiment We put a lot of effort into validating the products

as we go, and hope that our users can benefit from that by following those same protocols We realize that’s not always the case, but that’s where tech support comes in because oftentimes we’ve tested our antibodies under other conditions as well, and can help provide some insight as to the effect of say using milk or BSA to blot during the primary, or blot, or use 4% paraformaldehyde versus 3% during the fixation step When we provide our technical support, we should be helping you understand why an antibody didn’t work, if it didn’t work, or how to improve your results Sometimes that’s as simple as providing a separate protocol, but other times that involves a lengthy troubleshooting process We also have application experts In other words, when you call our tech support line, you’re not directed to somebody who then routes your call You’re actually directed to the people who validated the antibody by that application If you have an IHC question, you can talk to the IHC team If you have an ELISA question, you talk to the ELISA team We also offer a number of tutorials and application guides which should help with commonly asked questions We offer on-site training for some of our more difficult applications, such as our proteomics work We collaborate widely with investigators Not only investigators developing new methods, but investigators who are trying to use our antibody for a method that we haven’t validated the product We do this as a way of ensuring that our antibodies are handled and used in a way that’s consistent with our own practices, so that when other customers want to use the antibody in that method, they can benefit from previous customers’ work The materials that CST makes are tested for specificity, sensitivity, reactivity, consistency, stability, and manufacturing As I mentioned earlier, we manufacture 100% of our antibodies in-house Which means that everything that goes out the door has been tested by one of our production teams I’ll walk through that process in a minute The stability issue is a big one I think most people aren’t of is that rabbit monoclonal antibodies, especially recombinant rabbit monoclonal antibodies, are pretty stable However, before we send anything out the door, we store it at 37 degrees for one week to make sure that it performs as well before and after it went into those conditions Many of our products are shipped according to their stability, and that’s clearly indicated on the data sheet From a consistency perspective, we test every lot of a new product in every approved application before the new lot is released We make sure that the new lot looks just the same as the old lot when we release those products I’ll show you an example of that in a bit First, I want to walk you through a little bit of how we show data, and all the processes we go through to validate antibodies in the hopes that you can start using some of these in your own lab to make decisions on purchasing antibodies, but then also save yourself time and effort when you’re using an antibody in a new system This is a typical data sheet The same information’s available on our web page On this data sheet, we show information about the approved applications and the tested species that the antibody has been used in When we say approved application, that means that here are the applications that we’ve tested this antibody in, and the antibody has passed We don’t indicate any applications that have failed, or were not tested If you need that information you can call our tech support line, and in the future, we will make that information available to users We also indicate the species that the antibody works against Some antibodies, we’re able to test a large number of species Others, we’re only able to test a few By default, we always test in human, mouse and rat, and make that information available where possible If we list a species in parenthesis, it means that the antigen is 100% identical over that antigenic range We only list those species that are 100% identical We don’t make any assumptions We also list the storage conditions We list the applications, additional information about the applications, the dilution that the antibody was used, and the buffers Now one shortcoming that we realize is that many customers like to see the concentration of the antibody We don’t actually list that on our data sheets We are gonna start putting that on what we call our certificate of analysis, which is a lot-specific piece of information that will be attached to each product on the web page In the meantime, if you need that information, you can always call our tech support line and they will gladly give you the information We also indicate the specificity of the antibody, and this is especially relevant if there are family members against the target,

and so we will clearly state as best of our knowledge that whether an antibody cross-reacts within an isoform We also talk about the sensitivity, and give you the antigen information Finally by default, on all of our data sheets, where an antibody’s approved for Western blot, we will give a very quick piece of detail about the Western blotting conditions That includes the dilution, the blocking, and the primary conditions The reason for this is that some of our antibodies work really, really well in milk, and really, really poorly in BSA, and vice versa By default, we test all of our antibodies by Western blot using both milk and BSA as the primary diluent When we do that, we get really divergent results Now sometimes it doesn’t make any difference, but many times it makes a significant difference, especially for phospho-specific antibodies, and so we attempt to make that front and center, a piece of detail that you can easily refer to, to understand how to optimize your Western blot condition When using an antibody from any company, I encourage you to use a variety of different buffer conditions in order to optimize performance in the antibody Why do we go through all of this effort to make the antibodies that we do? Part of the reason is that we’ve listened to our customers Our teams travel widely We have over 250 active scientific collaborations as of today, and we hosts scientists all the time at our facility to come give talks We spend a lot of time talking to scientists to understand what it is they need, what it is they want, and how we can make their antibody use more seamless We spend a lot of our money reinvesting back in our antibody development pipeline I think that CST is unique in the fact that because we’re privately owned, we can do that It’s really great to be in an environment where we’re developing new technologies, we are investing heavily in new platforms for antibody validation, and we are investing in the people that validate those antibodies We take a pre-market/post-market approach to antibody selling That is that we engage scientists upfront to understand what products we need to develop, but then in the end, we also need to continue to support those products by engaging customers to understand, is something not working right? What new applications do we need to try to test this antibody for? How can we make this product better and meet new needs? Our antibody development process looks a lot like this, and I’m not gonna go through this in a whole lot of detail, but this is how we make a typical rabbit monoclonal antibody It starts with target selection We then design the project based on the application requirements, species requirements, et cetera We then immunize and bleed the animals We go through a round of application testing Again, the applications chosen at this point are dictated by the target of the antibody For example, if we made a PD-L1 antibody for Western blot only, it probably wouldn’t be that useful to investigators We make immunohistochemistry the primary driver for that reagent, and then we test that antibody by IHC upfront We then go through a series of molecular cloning steps to ensure that we’re capturing the antibody That goes through another round of application testings to make sure that the clone that we selected early on, again, is the clone after recombinant cloning There’s a set of final clone selections, yet another round of application testing, and then that’s when the product is developed and provided over to production to be made as a larger lot That lot is then made and then tested by an independent group Our production group and our development groups are independent, so the teams that are doing the lot testing are different from those that are doing all the development testing Every product is being tested by at least two different teams before it goes out the door Those teams have to agree on the quality of the product before we release it I’d like to say that everything that we take on becomes a product, but that’s not the case Almost half of our antibody projects end up failing at some point during the process We actually throw away just as many projects as we end up making in the products, which I think is a huge attrition rate, but it also goes to show how important our antibody principles and our validation steps are Out of interest in time, I’m gonna go through this very quickly In general, these are our antibody principles, our validation principles We determine specificity, sensitivity, consistency, we optimize the methods,

and we provide high-quality tech support What I’d like to do is drill down in a little bit more detail about a couple of these The first being sensitivity, the other being consistency I want to show you some examples now of how we use specificity to validate antibodies This table is designed to be an incomplete example of some of the methods that we use to determine specificity I want to call your attention specifically, the use of positive or negative tissues, or use of knockouts to make sure that antibody is specific By way of example, we’re using immunofluorescence to look in positive and negative cell lines to make sure is the antibody specific? Now again, this particular experiment just shows the antibody is specific by immunofluorescence Doesn’t say anything about how it performs by Western blot You have to do the same specificity analysis that you would do by Western blot that you do in this example We never just do one method We often use multiple methods to confirm specificity For immunofluorescence in particular, and then sometimes for Western blot, we’ll use subcellular localization as an indicator of specificity If you have a cytoplasmic protein, and it stains the nucleus then you probably don’t have a specific antibody We will also use self-fractionation to test antibodies by Western blot For the most part, anything that is not specific in localizations in one application, with that calls into then question results about the other applications, unless we can prove otherwise Using agonists, or inhibitors, or growth conditions to shift a protein around is also critically important to validating an antibody CST offers a wide array of fosto, or modification-specific antibodies, and in these cases, we very often use treatments in order to ensure to convince ourselves that the antibody is correct Again, we will also use positive and negative cell lines, but the treatment condition really gives us the right information that we’re looking for We often compare our products against those from other companies This was actually a product that I made, the top image of prostatic acid phosphatase, was tested both by Western blot, and by immunohistochemistry against the top competitor, and the oriol J-clone, which is used in a clinical diagnostic marker We will often screen our antibodies against these competitor to see where we stand Is ours better, worse, or the same as competitor antibody? Oftentimes we will only sell it if it’s an improvement on the market Finally, we’ll use other techniques, such in this case is using cell pellets to prescreen for immunohistochemistry What you’re looking at here is the CST antibody highlighted in blue, or a non-CST antibody, which is on the right-hand side Met positive cells in the top two boxes, or met negative cells in the bottom three boxes, using CST’s new anti-met antibody You can see that in our hands what we’re seeing is a nice, specific staining in the met positive cells, no staining in the met negative cells Whereas a competitor antibody stains equivalently, if not more so in the negative cells In terms of consistency, we go to great lengths to ensure that every lot of our product is exactly the same as the previous lot If you’re switching lots, you’re not suddenly gonna start getting new results This is just one of many examples that I have This is a rabbit polyclonal antibody It’s one of our best-selling products, number 9101 It’s against phospho erk What you can see in this image is that one of the very early lots that we made, this lot 10 rabbit polyclonal was released in 1999 I’m showing you examples of additional lots that were sold after that, that have since been retired Lots made in 2000, 2002, et cetera You can see that at least by Western blot, every single lot, it looks exactly the same as all the other lots We go through great lengths to make sure that our formulations are correct We would follow up also testing all of these lots by IF and by the error-proof applications to make sure that the product performs equivalently at exactly the same dilution in every single application for which the product was approved From a tech support question, as I’ve said many times, it’s very important to go through, if you’re having an issue with one of our products, please call, and somebody will help you You can also email or go to our website to submit a form I want to spend a couple seconds talking about what others are doing to address reproducibility

There have been a really great number of initiatives that have been proposed For example, there are now some reagent idea repositories, such as the RRID system that’s been proposed and being used by some journals There are also protocol sites I’m just citing an example of Protocols i.o which is being used to track reagents not only in labs, but in publications, and track the use of protocols I encourage you to check both of those out as a way of citing and reviewing protocols and reagents in the literature and in your own work A number of really good antibody search engines that become available, and again, ones that link to the vendor are probably best The ones that offer to sell you a product are probably just resellers A lot of these engines do a really nice job of tying a product to the number of citations, or end user data I find that to be really useful in terms of finding out which antibodies are most widely used, most highly published, or highly recommended by others Journals have started to make a flow change in terms of encouraging better citations and better user behavior by implementing a checklist, making sure that authors are actually listing the antibodies correctly in current use of proper methods I know Cell Press has launched STAR As I mentioned above, some journals are working with the RRID group to ensure that RRIDs are used to cite antibodies There are a number of individual journals in which the editors have come out to say, you need to be more careful about how you list and cite your antibodies The NIH has started a program of Rigor and Reproducibility in Research, and they’ve launched some efforts into finding ways to ensure greater reproducibility and greater accountability in research CiteAb, a group out of the UK launched a meeting earlier in 2016 to bring together scientists, and other vendors, and journals to talk about antibody validation Most importantly, the GBSI, which is again, the Global Biological Standards Institute recently held an antibody validation workshop as part of their reproducibility initiative They invited vendors, research scientists, journal editors, funding agencies, and antibody database experts, and search repository leaders to come together to discuss what’s wrong with antibody validation and what needs to be done to fix it Out of that meeting came a number of initiatives, and one that I’d like to highlight again, as I’ve mentioned, is the need for greater reagent or resource identification in the literature, et cetera This is where I think RRIDs makes a huge impact on tracking not only the reagents that are used in papers, but eventually other information, such as the lot, the dilution, things like that The vendors are beginning to work with the GBSI to adopt a set of validation standards Those are being worked on now, and so you’ll have to stay tuned for those Basically, the standards that they’re coming up with, if you’re familiar with the five pillars, workflow, that was published earlier this year in Nature Methods, or things that, to be honest, CST is already doing We still have some work to do, but for the most part, we feel that we’re already adhering to those validation standards We also make sure that we use detailed protocols That was another topic of discussion at the GBSI meeting We like to think that we’re ahead of the curve on that, but again, could do a better job We also, another topic that came out of the GBSI meeting was the need for vendors to have increased transparency, to show more data, to talk about where the antibody is derived from, so to make resellers more accountable for what they’re reselling, and to show more application-specific validation data Researchers, as I mentioned earlier, were tasked with finding ways to better train, or better mentor, especially the younger students in their lab, and to increase the use of the reagent identification resources that I mentioned earlier Until the working group comes up with guidelines, validation guidelines, I just want close by offering a few of my own advice based on experience in my work here on how to find a good antibody It starts with looking at the data, reviewing the data, both that provided by the vendor, any publications where the antibody is used,

albeit that that’s difficult to find, but the search tools now are becoming better at doing that Looking for how the antibody is validated Does the vendor show testing data or validation data? That’s what is really critical You need to look at the applications that you’re using because again, like I said, if there’s one thing that you need to take away from this, it’s that if an antibody works in one application, it doesn’t guarantee performance in another You need to look If you want to use an antibody for ChIP, was the antibody validated for ChIP? Does the data look like it’s quality data? Did the vendor, or the person who developed the antibody, use the correct gene targets to validate their ChIP? You need to look and review the data very, very clearly, and if the vendor doesn’t show it, you should call them and ask them You also want to check to make sure that the species reactivity is correct for the model that you’re using If a vendor sells an antibody, and it was raised against human protein, there’s no guarantee it will work for mouse You need to make sure that the vendor actually tested for mouse, and you can either look on their website, CST makes this clearly available, or you should be able to call and find out The vendor should provide protocols, especially application-specific protocols for how to use the antibody Preferably the dilution, or a very tight dilution range It kills me when people say to determine by end user, because that doesn’t make any sense You’re having to redo a bunch of validation data for the vendor I would suggest looking for vendors that not only offer specific protocols, but give you narrow ranges on how to use the antibody Now you’re gonna have to do some titration on your own end, especially for some applications, but for the most part, at least they can give you some guidance as to where to start and where to end If it’s important to you, you might want to look at the type of antibody Recombinant antibodies tend to be, at least rabbit and mouse recombinant antibodies tend to be much more reproducible than polyclonal antibodies If polyclonal’s the only things that’s available, you want to see about the lot to lot consistency Is this something that you’re gonna make a big discovery on one day, and find out that you can’t reproduce it six months from now because the vendor’s gonna switch to a new lot? You may also want to review purity in citations Find out where the antibody was used, how it was used, if the vendor didn’t approve the application you’re looking for, did somebody else use it that way? Can you find out any more information? Does the vendor offer technical support to be able to help you if you have a problem? All companies should direct you right to somebody who can give you true technical support, not just send you a free tube of antibody I would also like to say that some reviews are helpful, but I would use these with caution Not all product reviews on websites are accurate I prefer to use the ones in which somebody has actually supplied the data to go along with it I would use those with a grain of salt With that, I will close, and say that if anybody has any questions, I encourage you to either go ahead and email me directly, my email address is up on the screen, or if you go to www.CellSignal.com and browse the about us section of our top navigation, we have a page dedicated to our antibody validated principles Now and in the future, there will be a lot more detail about our validation processes, the transparency that we are going to implement, and some of the changes that we’re gonna make in the way that we do our antibody validation Like I said, we’re not perfect, but we are we think, right where we need to be in terms of high-quality product development and distribution With that, I will close and say thank you to everyone – [Robert] Well thank you so very much for that informative presentation Now let’s get started on the question and answer session I would like to remind our audience how to send a question You can submit questions by typing them in the Q an A box, which can be found by clicking on the green Q and A button at their lower left of the presentation window We’ll try to answer as many questions as we can Our first question For monoclonals, do you provide the epitope sequence information? Do you provide information about cross-relativity of epitopes overlap? Example, AB that binds to a protein epitope, versus another that binds to a same region when phospho (mumbling) – [Anthony] Yeah, so we started doing a lot more of that now I think as it’s become clear that post-translational modifications that occur very, very closely to each other can affect antibody binding If you’re trying to detect one post-translational modification, and there’s another nearby post-translational modification, it might change

We do For our antibodies that are raised against peptides, we are going to make those peptide sequences available on our website Currently, they are not We give you the information about the region around which the antibody was designed If you call our tech support line, they will provide that information to our academic customers The reason we don’t, is because there’s proprietary issues around the antigen design, and things like that But for the most part, to our academic researchers, this is not a problem The answer to the question is that we use a number of techniques to determine the specificity of antibodies that are near multiple PTMs The first is that we’ve been using for a while now, antibody arrays, in which we spot, sorry, peptide arrays, in which we spot a number of different peptides, each containing a different iteration of the post-translational modifications in that region We spot the peptides at varying concentrations, and then use the antibody to determine its specificity The other way we do this is to make either knockouts, so we’ve done this for a number of projects, where we’ll make serine to alanine mutations, or other mutations to try to see if the antibody still reacts with the protein when we mutate it, or otherwise modify the site Our antibodies we try to make as specific as possible, using as many methods as we possibly can In some cases, it’s very difficult, but we do our best, and we will provide this information if it’s needed – [Robert] Thank you What are some differences That you’ve experienced with recombinant antibodies and (mumbling) antibodies? – [Anthony] I’m sorry, could you repeat the question? – [Robert] Oh yes, I apologize What are some of the differences you’ve experienced with recombinant antibodies and (mumbling) antibodies? – [Anthony] Ah yes, good question Thank you Recombinant again, is all that we’ve done is taken the antibody and expressed it in a heterologous system What we find with antibodies derived from hybridomas, whether it’s a mouse or a rabbit, is that there is some lot to lot variability Mouse monoclonals in hybridoma are relatively stable, but rabbit monoclonals are much less so What we find when we developed the recombinant clone is that we get much greater consistency, the lot to lot performance goes way up, and we don’t risk actually losing the antibody due to a crash of the hybridoma We’re going in that direction because it makes our reagents a lot more consistent In terms of performance by application, we found that there is no recombinant antibodies, let me put it this way There’s no evidence to suggest the recombinant antibody performs any better than a traditional monoclonal antibody in any application You still have to validate a recombinant antibody You have to test every lot because there can be, depending on how you culture your cells, differences from lot to lot, which is why we make all of our recombinant antibodies under very carefully controlled circumstances For the most part, when you hear people say that recombinant antibodies are better, I’m not sure that there is evidence to support that Production and manufacture is easier, but they’re not necessarily any better than their counterparts Some people actually argue that polyclonals, because they recognize multiple motifs, actually give you greater performance in some applications I would buy that argument – [Robert] Thank you I use anti-mouse antibodies to study antigen in lizard kidneys Is this result valid, or no? – [Anthony] I’m sorry Could you repeat the question? – [Robert] Yes I use anti-mouse antibodies to study antigen in lizard kidneys Is the result valid, or no? – [Anthony] I don’t have enough experience to be able to answer that question, sorry – [Robert] Not a problem Thank you For antibodies that are not available, does CST welcome suggestions for antibody targets to be designed? – [Anthony] Absolutely I can’t bring up the website right now, but we do have a place on our website

where you can suggest a new target Or if you can contact your sales representative, the way that it works is that our sales representatives will take those suggestions and bring them to our development scientists, who will then consider them – [Robert] My question is related to the spec sheet in reagent use for validation In most of the CST One moment here I apologize Oh, I apologize The question is, how can we address (mumbling) the CST antibodies in R-01 application as a part of the scientific (mumbling)? – [Anthony] Yeah, so it is my understanding, that’s a good question It’s my understanding that the NIH office of Rigor and Reproducibility is engaging a new system in grants in which they’re more highly vetted My guess is that in the very near future, grants are gonna have to include some sort of validation data for some of the materials being used I don’t remember There was a conversation around this very issue, but I honestly don’t remember what the outcome of that discussion was Whether or not the vendor data would be sufficient I think the reason it wouldn’t be is that if the vendor hasn’t tested the antibody in the application and the cell liner tissue that you plan on using, that the vendor data would not be sufficient For example, if we show, or the vendor shows, reactivity in HeLa cells, and you’re using mouse brain samples, you may actually have to show the antibody validates in mouse brain, instead of providing the vendor data, unless the vendor data has done that experiment for you – [Robert] Can you also write the IHC antibody or test on automated system, and maybe share the protocol as well? – [Anthony] That’s another terrific question We historically have not used auto stainers, only because we don’t have access to one We are in the next few weeks, about to bring in an auto stainer to CST, and we’ll start to validate some of our IHC antibodies for use in an auto stainer, and will make that application and the protocols available on the website, and upon request, if anybody needs it – [Robert] Are there any instances where the choice of the secondary has consequences on the quality of the result? Do you always use the same one, and do you do any testing with multiple antibodies? – [Anthony] That’s another good question At CST, as we’re doing the testing process, we use our own secondary antibodies so that we don’t have to buy them from anybody Does the secondary antibody make a difference? Absolutely The dilution that you use as secondary antibody, especially for Western blotting, the dilution makes a huge impact on the signal to noise ratio, on how dirty the background is, and we do occasionally, we’ll test our competitor’s secondary antibodies We typically don’t make that information available We do it for our own purposes In general, yes The choice of a secondary antibody, and how you use it, makes a huge difference, and I would recommend that the first time you use a new antibody, you should also do a titration with your secondary to make sure that you’re using that at an optimal concentration, as well – [Robert] Can you talk briefly about the role of buffers, and its importance in IHC, and link it with the retrieval solutions, please? – [Anthony] Absolutely We find very often that the difference between an antibody working or not working, is entirely due to a number of factors The fixation conditions, the age of the fixative, so how fresh your buffers are, the antigen retrieval methods, and the conditions under which you do antigen retrieval All of these have a huge impact We’ve worked with a number of large clinical labs, that when they’re doing clinical trials often see wide variability in the antibody staining pattern because there’s no consistency over how the antibody,

or how the material is fixed, the age, how long it sat in the fixative We find these to be really, really critical components in how an antibody performs in an IHC assay IHC and validation by IHC is almost kind of its own universe People that do a lot of IHC will tell you that you really need to validate an IHC antibody from scratch in your system We do as much as we possibly can to facilitate that One thing that we do do, if you’re doing IHC, is that we do make available control slides that will allow you to, that we have tested internally with our own antibody, so that you can take those control slides and run them through your own protocol with our antibody to make sure that our antibody is still working I encourage you to use vendors that do that kind of thing We also provide lysate controls for Western blots, and things like that, as a way of controlling for sample preparation variability that occurs – [Robert] Next question now is related to the spreadsheet and the agents used for validation, and most of the CST spreadsheet, only CST reagents mentioned, such as dilutants and detection reagents Do you use most commonly used reagents such as (mumbling) PDS and dilutant, or different secondary reagents? – [Anthony] We use pretty common buffers A few of our antibodies work in buffers that are specific for them, and we typically call them out On our website, on each product page, there’s an accordion that you can open that talks about the protocols Within that area, you can actually find out more details about the buffer compositions that are used For some of our reagents where that’s not indicated, I would call our tech support line, and they can give you that information None of the buffers are trade secret situations, so we’re pretty open with that – [Robert] Thank you In what type of malignant celled lines are recommended for validating antibodies specifically using CRISPR? – [Anthony] Oh, yeah That’s a good question We typically like to use a cell line We use many of the known genomic proteomic databases to find cells that are predicted to express high levels of our protein of interest Ideally, we like to use those cell lines to knock down using CRISPR The reason for that is that some of the model systems that are available out there, such as the CRISPR line from Haplogen, those cell lines, which we’ve used in-house, don’t express all proteins You’re gonna get a negative signal, and should get a negative signal, in both the wild type and the knockdown cells We prefer to use human cell lines that are predicted to express decent levels of the protein Maybe not super high, because then it makes the CRISPR harder, but that we can confirm by Western blot, express a good amount of the protein of interest Then we can use CRISPR to knock that down, down or out completely – [Robert] Doctor, do you have any final comments for today? – [Anthony] I don’t I would just say that if there are any questions, I’m more than happy to answer them My new job here at CST is to keep tabs on the reproducibility initiative, and to make sure that we’re supporting our customers in any way possible to make sure that their work is reproducible If anybody has any questions, feel free to email me If there are questions after the webinar, I would encourage the same – [Robert] Thank you I’d also like to thank our sponsor, Cell Signaling Technology, for making today’s educational webcast possible Today’s webcast will be available for on-demand viewing through April of 2017 You’ll receive an email from LabRoots alerting you when this webcast will be available for replay, or invite you to forward that announcement to your colleagues who may have missed today’s live event Until next time, goodbye