Best Practices Guide to Sample Preparation

Table of Contents

Letter from the Editor
Index of Experts
Experts Speak:
Betsy Bove
Jan Kieleczawa
Kevin Knudtson
Suzanne Perry
Sunny Tam
Michael Zianni

Download the PDF version here

Letter from the Editor

How do you get the best possible results? What is it that makes them irrefutable? In this supplement to Genome Technology, we bring you the thoughts and perspectives of those hard-working core lab facility managers on sample preparation and quality control. We asked an array of people from different facilities how they maintain their high quality standards — and what their customers can do to help. Their responses, no matter if they hailed from sequencing or proteomics or clinical molecular genetics, boiled down to, "Do your homework." It'll be more painful — and more expensive — to find out after sequencing your clones that your insert wasn't there than doing a quick gel run a week or so beforehand. Flip through the rest of the pages to see what else these core lab experts have to say that could save you time and money on that future experiment.

— Ciara Curtin

Index of Experts

Many thanks to each of the people below for taking time out of their busy schedules to speak with GT to bring you this supplement.

Betsy Bove
Laboratory Manager
Clinical Molecular Genetics Laboratory
Fox Chase Cancer Center

Jan Kieleczawa
Biological Technologies
Wyeth Research

Kevin Knudtson
DNA Facility
University of Iowa

Sunny Tam
Director of Proteomic Fractionation Group
Research Associate Professor
University of Massachusetts
Medical School

Suzanne Perry
Manager and Research Scientist
Proteomics Core Facility
University of British Columbia

Michael Zianni
Manager/ Senior Research Associate
Plant-Microbe Genomics Facility

Betsy Bove
Betsy Bove manages the clinical medical genetics laboratory at the Fox Chase Cancer Center in Philadelphia

Genome Technology: How much of a role do PIs play in sample preparation? Do you train them?

Betsy Bove: We're a clinical testing laboratory — CAP and CLIA-approved — so only qualified techs handle samples, and sampling is done in our outpatient clinic facility. The sampling is not actually done in our laboratory, but we do have protocols that dictate, through CAP, sample preparation. We know that the clinical lab is an arm of our pathology group that is also under the same umbrella of CAP, so they supposedly abide by these quality controls.

GT: What steps do you take to ensure good sample preparation?

BB: When we receive the stuff, then we have quality control protocol standards of receipt so that the sample has to be adequately labeled. It has to be of proper quantity and quality. There can be no evidence of co-mingling of samples so that the sample may have been tampered with. Those are all dictated in quality control and quality assessment, again dictated by CAP. We're in good stead with our clinical laboratory; it's a floor and half away from us so if there's an issue, we call and they are top-notch, I have to say, so any issues are quickly resolved.

GT: How do you validate new technologies, methodologies, or machines?

BB: We use a validation procedure that is, again, dictated by CAP [and] also by our department of pathology. Our MD pathologist, who is the associate head of the department, oversees any kind of new validation which is held for either technology, methodology, or any machine. We just test a panel of samples that have formerly been tested by an already approved assay and compare those results to the new technology, methodology, or machine being used and be sure that the outcomes are reproducible, sensitive, and specific. I think that's pretty standard protocol, actually.

GT: How do you assess or maintain your quality standards?

BB: We have dictated a program of quality control and quality assessment which dictates the standards for all reagents, machinery, actually technology prowess as well. So again, they are instituted in these two main protocols of quality control and quality management so they are systems that include weekly or monthly assessment of machinery, reagents, data quality, completeness of data, gel quality that we review weekly and monthly.

GT: How do you impart your quality ideals to your customers?

BB: Our customers meaning clinicians. Well, they know that we're a CAP-approved laboratory under the arm of our department of pathology, so they are already in good stead there. They assume that we adhere to the same standard, which we hope to. And any other questions as far as new assays, the clinician will call, personally, the laboratory and we're usually able to answer questions.

GT: How can your customers help you maintain high standards of quality?

BB: Going through the correct protocol and both recruiting a sample for testing and in managing the transport of the stuff from the patient to the laboratory — and that's really harder to achieve than I had anticipated. There are many ways to get the sample to the laboratory, but we try to keep it routine. They are very interested in trying to do that but molecular pathology is fairly new across the globe, but especially here at Fox Chase. So it's a totally new endeavor. But the clinicians, they are happy and anxious to have the service so they are coming right along.

Jan Kieleczawa
Jan Kieleczawa is the laboratory manager of the DNA sequencing group at Wyeth Research.

Genome Technology: How much of a role do PIs play in sample preparation? Do you train them?

Jan Kieleczawa: We are a DNA sequencing core facility. We do practical DNA sequencing and oligo synthesis only. So what we do, we recommend best procedures from the cDNA sequencing point of view. Awhile ago, I tested a bunch of different DNA preps using six different commercial kits and then based on this study, we have our favorites which will work especially good when the templates are difficult, something unusual about the templates. The most favorite is the Edge Biosystem method for DNA prep, for example, and a Qiagen. Based on our experience, we do recommend certain DNA prep methods. As well as a clean-up procedure for PCR fragment for somebody who does PCRs or remove primers, we also have a broad experience in this. We know what works best, what works so-so. Definitely yes, we have a role — whether people listen or not. Sometimes you simply use what you have and hope for the best.

GT: What steps do you take to ensure good sample preparation?

JK: First of all, I will follow recommendations and then I would optimize. When a kit comes, it is meant for a very broad application. If you want to do some specific application, I would simply optimize for my specific use. I would spend a lot of time doing all kinds of optimization. Anything related to sequencing, you can broadly break the sequencing process in over 10 different smaller steps and I have optimized every single step of this: DNA prep, quantitation, sequencing, cycling protocols, clean-up of sequencing reactions. [Every] single step, it was optimized in our laboratory, so we know we have the best condition to get the best data. I would start with the recommended protocol and then tweak it.

GT: How do you validate new technologies, methodologies, or machines?

JK: Same way, I would do it.

GT: How do you assess or maintain your quality standards?

JK: Once again, for us it is relatively easy because the measure of our quality is the read lengths in the DNA sequence, as measured by the Q-value. We have a database. We dump all the data on a server and you can get a quality score for the day, for the month, for the week, and this way you can say, "OK, something is wrong with the machine." Not only do you have a numeric value, but you also have a visual value. Something that is very good is green, something that is black. You can right away spot, for example, trends that allow you to know that something is wrong with it. You can really visually assess that everything's OK or not OK.

GT: How do you impart your quality ideals to your customers?

JK: It's a pretty simple principle. The idea is: "Garbage in, garbage out." If you do not give me good DNA, I cannot do anything; that's the basics. In the sequencing, at least, the newest generation machines are pretty forgiving. A long time ago when I was a lab technologist, everything had to be really pretty much in a very finite, narrow range, DNA concentration and this and this. These ones are not that strict, much better. Even DNA preps that used to not work on this lab technology would be OK, not the best, but they would be OK. In general, you have to ask for the best quality of extracting material to get the best possible results. We will ask in advance, "Do you expect that your DNA could be difficult?" "Yes." "You have the best chance if you use this method to prepare this DNA this way." So this way people will have a better chance to do it. We will do of course whatever they give us and try to do the best but if it doesn't work, we would recommend something different: "Try this one, you have a better chance to do it."

GT: How can your customers help you maintain high standards of quality?

JK: They want the best, all the time. We used to give them about 1,000 bases and if it's less than 800, they complain. We have to be on our toes all the time. What it really means for us, we have to make sure that our instrumentation always has a PM, preventative maintenance, that we have all the calibrations on time and everything's in top quality, fresh buffer, whatever needs to be done always has to be the best quality. Being in a pharma company, we are not a pinch-penny, if we can get data tomorrow instead of a week later. But it's not something overriding that you have to because if it costs one cent more that competitor A, you're out of luck, it's better that you get data as quickly as possible, as best quality as possible. It's more important than saving a couple of dollars on our samples. That's, in principle, how we work. We will recommend, don't use as a scanning tool. For example, somebody will have a hundred clones. Typically the way they are supposed to do, I will run this on the gel and discard this one which has, for example, no insert or they are low-quality, low concentration, and give you only those which are good — but they usually don't. They give us everything and they will say, "OK, you tell me which sample is good." So actually in a sense, we act as someone to say a sample is good and which is not good, not the way it is supposed to be. But anyway, they would help themselves if they would do upfront work. Sometimes you have one or two clones, people will do it. If they have a hundred, sometimes it is much cheaper for us to do it; they would have to prepare their gel and run it and it's better that we do that gel than they do that gel. It doesn't happen enough. Sometimes we get a relapse because people give us 10 gig clone and then we do hundreds of reactions and there's no insert there. They could have done this upfront, run the gels and say, "OK, I have no insert, what's the point of sequencing?" We still have to spend time, look at the data, and tell them thatsomething's not right. But everywhere people are very conscious of that and they will do everything to give us the best quality.

Kevin Knudtson
Kevin Knudtson runs the University of Iowa DNA Facility.

GT: How much of a role do PIs play in sample preparation? Do you train them?

KK: In our facility they play the primary role in sample preparation. We do not prepare samples for investigators that use our services. The reason behind that is, we deal with a lot of different samples — samples derived from human tissue, mouse tissue, flies, plants — and there are techniques that are optimized for these various tissues, and the lab has more skills in effectively purifying the RNA then we might. For those labs that are not able to or don't have any idea, we do train and provide some advice, general guidelines of how to acquire a quality sample.

GT: What steps do you take to ensure good sample preparation?

KK: We work with investigators and they have generated some RNA or DNA, mainly RNA, and they want to know, "Is this RNA sample of quality to maybe go on to the protocols, say microarrays, for example? Is it of good quality?" We work with investigators in assessing the quality with a bioanalyzer and we can look at quantities and get some kind of assessment of how clean things are with a spectrophotomer, the NanoDrop, doing a UV scan. We are working with investigators to help them assess quality of the samples that they generate. Also, they'll maybe run the samples out on a gel to see the DNA sample and want to make sure it's high molecular weight DNA.

GT: How do you validate new technologies, methodologies, or machines?

KK: We try to run control samples to assess the accuracy and sensitivity of the instrument. We want to look at the noise, sample, reproducibility of the results. We are running replicates within the same run, if possible, then also the reproducibility from day to day. These are essentially run with controls, both positive and negative controls.

GT: How do you assess or maintain your quality standards?

KK: We have quality control standards for many of our instruments. For each, say microarray, there are certain benchmarks you need to achieve with a qPCR or there are certain things we can look at. We do run quality control standards. We make sure the instruments are performing as expected. We quality control our instruments to make sure that the output is what it is supposed to be. We're running standards on our instruments continuously. We run controls and standards on our instruments as well to make sure we're obtaining expected results. We do have standards that we do run on a consistent basis. We just want to make sure the machines are going to give reproducible data from run to run.

GT: How do you impart your quality ideals to your customers?

KK: When they start an experiment, we like to talk with the investigators and try to understand what question they are trying to ask by using the technologies in our cores. Then, we want to make sure that when — this kind of gets back to what steps do you take to ensure good sample preparation. Well, there is the quality of the sample itself, but a good sample also is biologically relevant.

This is where we want to work with investigators to make sure that when they are collecting their samples that they are collecting samples that will allow them to answer the biological question they are trying to ask. Some of the things that we look at are [what] we want them to be, consistency is key: using similar protocol to prepare all the samples in a particular study. If you use different protocols to prepare your samples on different times and dates, you're going to get a different answer with each different method you used to prepare those samples. The other thing is, you can pretty much tell them, "If you don't run good controls or you don't do a good experiment" — and these are part of quality ideals — "you're not going to get very useful data and you've probably wasted your money." That usually has an impact. The other thing is, "If you go to publish this and a reviewer sees that you really didn't design and run a good experiment, you're not going to be able to publish your data." That also impacts the customers as well. If they want to not waste their money or get their data published they should think about running a quality experiment.

GT: How can your customers help you maintain high standards of quality?

KK: Definitely they need to have a well-designed, well-controlled experiment. We'll work with that. Part of our talk with investigators is, "Let's design this experiment right, with an appropriate number of replicates, appropriate number of samples, appropriate sample collection, and we'll include the appropriate controls that will allow you to analyze and get useful information from this experiment."

Suzanne Perry
Suzanne Perry manages the proteomics core facility of the Michael Smith Laboratories at the University of British Columbia.

Genome Technology: How much of a role do PIs play in sample preparation? Do you train them?

Suzanne Perry: Well, I don't know if I would necessarily train a PI. What I usually do is to encourage them to come in with their sample directly — or if it's their postdoctoral fellows or research associates, lab manager, graduate student — I invite them to come and have a meeting with me. We talk about the nature of their sample, what their goals are in terms of mass spectrometry and then what their current status is, their work flow, so I can sort of slot in where they need to start looking at the requirements for their sample to be successful with mass spectrometry and we go from there. Some of them are much more involved in proteomics, some are fairly deeply involved, and some are not, just sort of starting out in proteomics. It's all over the ballpark when it comes to where they are in terms of sample prep.

GT: What steps do you take to ensure good sample preparation?

SP: I'm very clear on the products that we know work with mass spectrometry and which ones we will not accept. We have a Web page that we list compatible buffers and detergents, what
concentrations we can work with, which ones exceed our limitations and what the steps would be that we would have to charge them if they were to give us samples that were in a state that we would need to do buffer exchange or desalting. We also try to guide them through, if they decide to do the manipulations themselves in the lab, the necessity to avoid keratin contamination because that seems to be the biggest problem when we are doing high-throughput, high-sensitivity spectrometry. We just get tons of their hair, their sweater, their dog's fur in their sample. I don't think people realize although it's not a sterile technique and autoclaving doesn't kill keratin. Keratin isn't a live thing, it's a protein. That seems to be the problem. The way I talk to them in very simple terms about the necessity to really think about it as if they were doing sterile technique. The keratin won't be completely eliminated but reduced to a level where it won't confound a sample analysis, won't be competing.

GT: How do you validate new technologies, methodologies, or machines?

SP: Because we're a university core facility, I don't feel that validation in terms of what a commercial laboratory or a clinical setting might find important because we don't have to report to any of those bodies that might cast a critical eye on how we might run things. Our beauty is in the fact that we can push the envelope on our technology and do things a little bit differently to try to extract or bend or convolute things to get data that normally isn't available to a standard operating procedure. We're able to finesse things. We don't necessarily validate things but what we do do is before we get a piece of equipment, we will always research what we need in our workflow based on the current sample submission, where are we bottlenecking, where are we sort of not getting the sensitivity or the purity we need to carry on. We research very critically what we need, so by the time we actually get the equipment, we know pretty much what we are going to be getting in terms of bigger, better, faster, stronger, more sensitive from the types of things that we're getting. Then we compare it to past samples in terms of performance. Like where we might have gotten picomolar or centimolar sensitivity, we'll now get attomolar. Or if we had so many protein hits in a standard digest of a known protein, and maybe 27 percent coverage of a known protein, we might get a much greater percent of protein coverage at a much lower sensitivity, better resolution. If we do complex samples, sometimes we'll run it on lesser machine and then a more sophisticated machine and we'll see an increase with the same raw data in hits in terms of how many proteins were identified in a complex mixture on one machine versus another.

GT: How do you assess or maintain your quality standards?

SP: We have service contracts on all our machines. We subscribe to a very strict schedule of preventative maintenance that is actually put out by the company from whom we buy the service contract. My labs are really, really clean. In fact, service engineers have come in and said, "Oh my goodness, this is one of the cleanest labs I've worked in." I have really quality people who feel very strongly about the work that they do. I work with them and give a sense of autonomy so they have a sense of pride in their work. There are lots of other labs that can do the same thing we do, so it's important to us to give really consistent data. We take pride in the fact that we are cited in articles. Word of mouth is pretty much how we're found because people are pleased with their results and the consultation. I think that's the key factors in, at least, in my lab.

GT: How do you impart your quality ideals to your customers?

SP: I think the first thing you can do is be candid with what you can provide for them and, of course, being very patient with their questions, being very thorough in the type of information that you ask them for — and once you get that information, you have to be very judicious in telling them what you can do. It's not the Easy Button. Certainly we exceed most people's expectations. We don't undersell what we can do but we don't want to give people false hope. We do a huge amount of consultation and hand-holding, as it were. We will tell them, you must use this gel or you must use this stain. If you were to use this stain, it's not as sensitive. We are good at giving out samples if they don't currently have them in the lab. We pride ourselves on being a resource. We are always available for consultation. We don't charge for that. They spend a lot of time saying "Oh, no, my gel didn't work" and I'm like, "All right, bring it over." So, a lot of hand-holding.

GT: How can your customers help you maintain high standards of quality?

SP: There's a small percentage that drive me nuts. They come in and we call it "mouse in a blender." They'll say, "I want to know all the proteins in this sample." It's obvious to me that they haven't really understood the concept of proteomics, either top-down or bottom-up. They haven't spent the time to research what other people have already done in their field. They basically want me to do the project for them. That's not why a core facility is there. I like to think of ourselves as a station on a train track towards the goal, which is identification or posttranslational modification quantitation, whatever it is, and we're a vehicle for getting there — but we're not the whole engine. People need to do their research first. They need to come prepared to say, "I don't really know much about mass spec, but I know this system I am working with and I know that this has been done so far. Here's some papers." I get samples from everywhere from food science to drug discovery to aboriginal studies. I get all kinds of samples; I can't research samples. I can't know everything about everybody's protein in every system. They have to come armed with papers and amino acid sequences. If they want to do phosphopeptide analysis, they can't have every single serine, threonine, pyrosine labeled and expect me to find it quantitiatively. It has to be more specific. I think it's just the nature of the beast up here. We've gotten into the proteomics game a little slower than the rest of North America. That's a little bit frustrating at that point. It's not insurmountable. I think people have to sort of ratchet down their expectations.

Sunny Tam
Sunny Tam directs the Proteomic Fractionation Group at the University of Massachusetts Medical School.

Genome Technology: How much of a role do PIs play in sample preparation? Do you train them?

Sunny Tam: It's case by case. I think most of the PIs would not come here. But I do interact with them to make sure study design, any particular issues, precautions to their samples. It's case by case because everyone's materials are unique — there's no one-size-fits-all. In terms of the students, I train postdocs, graduate students, MD/PhDs, even undergraduate students in my lab on some of these techniques. It depends on what people are interested in. Sometimes they would just need the sample to get processed. We return to them in a format to take over as soon as possible. It's really not much teaching involved there.

GT: What steps do you take to ensure good sample preparation?

ST: Again, I think it depends on the question they ask. For instance, if they want to run a good 2-D gel, then in terms of what they give us, it's very critical. If they have a lot of phosphate — people do a lot of a cell culture washing with PBS, those types of things, we tell them to switch to Tris-buffer saline, for instance, so that we have less conductive ions in there. It really, I think, depends on what is the next step, could go into some liquid chromatography so it could be proteins that they over-express. Sometimes we have to do certain buffer exchange. So kind of things we have to be careful. The take-home message is, it depends on what question they are going to answer and second is, what kind of techniques [are] going to be followed from their samples. We have to make that transition in terms of the amount of materials, the type of buffer they should be in. All those things need to discussed ahead of time.

GT: How do you validate new technologies, methodologies, or machines?

ST: Well, you could always take samples that are less important and validate that with those types of materials. For instance, we have a pressure cycling equipment here. We just took some animal livers that's part of an old study to generate results, compare with before and after of the technology. Or sometimes, I think, when the technology is more well-validated, like we are working with the Protein Forest DPC chips to separate proteins by their PI range, that technology is a little more mature and we are more comfortable with it. We can test on biological fluids that [are] part of a study. We can help to advance an external technology by validating that in increment in terms of how comfortable we are with that technology and all the technologies, we could always find a place in our workflow because our workflows are not rigid. It's really question-specific and sample-specific. We just want to take advantage of the most cost-effective and most informative technologies.

GT: How do you assess or maintain your quality standards?

ST: In terms of quality standards, it depends on what methods or protocol you are using. If it's 2D gel, we always look at it in terms of conductivity. Ions can affect the PI separations. So that's one QC in terms of the amount of protein that. We oftentimes use at least two assays for validating or QC materials. You could do a Bradford-Lowry or BCA to look at protein concentration, but also double-check with a 1D gel to make sure the protein load is similar to a standard. Always, I think, a couple of methods to double-check things.

GT: How do you impart your quality ideals to your customers?

ST: People look at the end results and they are happy with it. Part of the thing is the end also justifies the means. All these QC are very important: "Garbage in, garbage out." If you're using an antibody that is not immunoprecipitate, it's no use to anyone.

GT: How can your customers help you maintain high standards of quality?

ST: They work with us and also they expect sometimes things never work out the first time anyway. Just to be able to work with us scientifically, professionally, and do things in steps and make adjustments. For instance, we have done some 2D gel, some blots for a group, and we have to work with them to optimize the blocking of the membranes, those types of things, very interactive. They appreciate, I think, the efforts. Also, expect the unexpected.

Michael Zianni
Mike Zianni manages the plant-microbe genomics facility at Ohio State University.

GT: How much of a role do PIs play in sample preparation? Do you train them?

MZ: They play all the role, they play all the role. But we do give them advice. We do not give them detailed instructions but we do provide advice, general approaches and methods, and what has worked best given our past experiences here. We've been at this since 2000.

GT: What steps do you take to ensure good sample preparation?

MZ: The bulk of our samples are coming through DNA sequencing so it would be prohibitive, time-wise, for us to validate them — and difficult, given sample size. Typically, it would use too much sample.

GT: How do you validate new technologies, methodologies, or machines?

MZ: What we typically do is we seek out the vendors. Usually vendors have some sort of standards associated with the products whether it be protein samples, DNA samples, and so forth. We typically run those first and we use their conditions but also we'll will use alternate conditions, typically less amounts because we find that most vendors are highly motivated for us to buy their products. We can use less a lot of times and still get reasonable results. Once that works we can obtain what we need from their internal standards, we then seek clients that have known in the past to do good quality work, have good samples, and work well and communicate well with us. Then, we do their sample for free. We'll actually pretend it's a real client and do their results and analyze their samples at no cost to them and they get the results back. Sometimes it's with previous work they've done for confirmation and sometimes it's actually novel work and they'll take the chance that it works well. We typically like to see control reactions from previous work as well. Then we open the service up for a period of time under limited costs to make sure the equipment is validating and working well for a broader range of clientele. That's typically one to three months depending on the volume of samples and type of research done. After that, once we are convinced everything's working well, we go forward with it. What I didn't mention was oftentimes we try to get instrumentation in as demo units or go see other facilities, too. Our economic restraints are we have to recover all our costs: salaries, service contracts, supplies, through the fees we charge. We don't have an opportunity to throw money around. We're pretty constrained, so we try and be very judicious and wise in adoption of the technology. I also ask salespeople for references, other facilities or labs that have used their technology. We are not in a position to adopt absolute brand-new technology first time it comes off the shelf. We don't have that luxury here. We try and stay current, don't get me wrong.

GT: How do you assess or maintain your quality standards?

MZ: The first thing to do is we follow maintenance protocols quite strictly. I would say actually to the letter, but we do our best. In some cases we exceed the manufacturer's recommended maintenance. We value our instrumentation highly and make sure it stays in working order. We also periodically run control reactions as well. For our main sequencing service we run positive control reactions with every dataset and we also periodically run control reaction for all the other instrumentation as well. There is a constant monitoring of all the data that comes off the units manually.

GT: How do you impart your quality ideals to your customers?

MZ: We offer flat services here. I'm talking about genotyping, microarrays, proteomics, the whole prep. What we typically do, we require an initial consultation with all clients to let them know what we can do, what we can't do, what the guidelines are, presumably what's the most critical areas and where there's room for changes or modifications. We show them good results and we also show them poor results. We basically say, "We've got a lot of experience using this instrumentation and technologies, and it's in your best interest to follow this advice." It's difficult to make them do absolutely what we want. Ultimately, it's their research, is our attitude here. It's their results and they are ultimately responsible for it. We try and communicate well upfront and let them know what the expectations are and what they are not. I do not actively publish and make available to everybody our standards and show what we can do but I will show those to potential clients at the initial consultation phase.

GT: How can your customers help you maintain high standards of quality?

MZ: What we typical request is that they do the work upfront, i.e. they do some quality control on their own as well. We actually like to see that. If they don't, we don't deny them service. But if they've got it, we ask them to bring in preliminary data. So for example, for real-time PCR, it's tricky assessing what you are actually amplifying and measuring, so I typically buy generic primers at really cheap cost, do the PCR and sequencing the product and make sure you're amplifying what you [want] because PCR generates weird stuff all the time. And then go ahead and drop several hundred dollars on a probe set for doing real-time with all the standards and triplicates because you're going to put a lot of time and effort into it. Or, for proteomics, we like to see actual protein measurements on the samples for quantitation. We ask them to report the kit they used because there's a lot of discrepancies depending upon the extraction method in the kits, there's a lot of incompatibilities there that can give you false positive results. We also like them to do a 1D protein gel first to see that at least they have something in there before we proceed with the two-dimensional gel electrophoresis which, again, represents a week's worth of work, a lot of money, and then hours and hours of analysis afterward. We try to impress upon them that it's worthwhile to do, spend half a day, a couple of hours doing some quick initial test to confirm things.