Transcripts of the Attorney General's Initiative on DNA Laboratory Backlogs (AGID-LAB) Working Group

Monday, October 21, 2002

THE SOUTH AFRICA FORENSICS LAB AND PROCESSING

MR. VAN NIEKERK: What I would like to show you now is just a couple of pictures - well, quite a few pictures of the lab system that we have installed or that we are building and installing in South Africa, and I would like to use these pictures as a kickoff board to give you an indication of where we plan to go forward with the automation of crime samples.

The lab that we are currently working on, which is close to completion - the developmental validations of it has been done, the formal integration of the system must be done then, and our internal validations will take place before it's commissioned by the first of March of next year. This lab has been designed for the automated processing of offenders' samples.

There are three parts of this lab. The first part is the sample preparation lab where Marshal cassettes are being loaded into trays. If you will just excuse me, I'm going to dig in a box and I'll show you if you don't mind that. Just to add something to the voice, the first part of this lab is where these cassettes - I'll show them to you later on - are being added to a tray system like this, which is bar coded, which is then fit into the automated system, and after the handling by the lab technician of this cassette by putting it into the tray, after that until the time when we receive a result back from the system there is no human intervention anymore.

The preamplification lab basically takes care of the DNA purification, the punching of FTA paper, the PCR setup, and the amplification and detection lab takes care of DNA amplification and capillary electrophoresis for subsequent fragment analysis.

This is just an overview of what the lab looks like. We're completely aware of the fact that the robots - if I say "robots," I'm talking about articulated robot arms - is becoming a very well-known sight in labs all over the world. What does make this lab unique to a certain extent is that we've got the sample pit or the PCR setup lab and the DNA analysis lab next to one another only with an air lock in between. The physical spaces are further reinforced by means of differential pressures in the lab, so that if there is some kind of breaking in integrity of the lab in a wall, if a window is smashed out or whatever, then it will only flow in one direction and not allow contamination to take place towards the side where you still have the unidentified samples.

This just gives you another view of the lab. This view is taken from the preamplification lab, and if you look through the window, you look straight into the amplification detection lab. Next to it, just to orient you, this machine standing over here is an ABI 3100 genetic analyzer and then the track loading towards the air lock through which samples are received into the amplification and detection lab.

So just to show you these couple of pictures that I promised you, if we go through the process flow - I'm just going to mention this now. As we get to each stage I'll just pause and tell you more about that. The process flow in this lab is we have to log our samples into the system via LIMS. Then we need to have DNA purified - the FTA paper must be dried after it has been washed. We then punch the samples into PCR plates. We do PCR normalization and setup in a robotic liquid handler. The sample is then sent through an air lock to the next lab where DNA amplification takes case, the capillary electrophoresis setup takes place, and fragment analysis in a 16 capillary electrophoresis machine takes place.

Then afterwards just as a conclusion to this I would like to just to give you an idea of how we go about doing hitpicking, as we call it. If, for instance, we have a failed PCR or something like that, we can always go back and retrieve the sample and resend it through the system either coming straight from the blood sample or from samples stored in an automated storage and retrieval cold storage facility.

The first slide that I would like to show you is where we have our Marshal cassettes washed. This is a deviation from the normal way in which FTA paper is being processed. Normally what you do just in 30 seconds - normally what you want to do when you deal with FTA paper is you want to punch out a piece of blood stained paper, dispense that into a reaction tube, and then add the purification reagent to that, vortex the sample to wash out all the cell debris and PCR inhibitors and whatnot, and then you would carry on adding your PCR mix and so on to that after the sample has been washed.

What we do is we just switch over two steps. We first wash our sample and then punch, and the cassette has been designed to facilitate exactly that. So what we do is we have the robot arm place the tray containing the cassettes onto a custom designed and built washing station which facilitates the vacuum filtration of purification reagent through the paper and under vacuum it's just flushed away. From the top it looks roughly speaking like this.

What you see over there, those white dots next to the brown dot, the brown dot is where blood is spotted onto the FTA paper and the white dots are where blood is in the process of being washed away or the hemoglobin and debris is being washed away in this vacuum filtration system.

The drying is obviously essential because you sit with a soaked piece of paper after you've washed it, and you really don't want to exert the kind of force that we do in the system on a piece of paper. What you basically will end up with is shattered DNA through hydrostatic shock or something like that. So this machine, again it's a home brewed kind of thing, home designed. It's a drying oven incubator, if you will, which upgrades at 60 degrees centigrade, and it has got a rotating disk on which the trays are being placed, and 44 minutes later we take the tray out, and the sample is dry and ready to be punched.

This is just a view from one of the access doors into the drying oven showing you the disk with all the trays containing Marshal cassettes stacked on the work surface.

What I also should mention to you is that when these trays are loaded into the system, they're loaded into a magazine 24 trays at a time each containing four cassettes, which equates to 96 samples. Soevery time we add or we load a batch of samples it fills up a 96 PCR at a time.

I don't know if you can see this, but I tried to take these pictures just so you can see this. On the end of the robot arm there is a bar code scanner, and every single step along the way, whatever lab way we are using, the lab way is bar coded by means of that sample. Traceability is maintained by scanning every single step along the way. There the robot arm just goes ahead and picks up one of the trays. It either puts it in or takes it out, whichever way you will.

The next step is the punching of the samples, and if I say "punching," it basically boils down to the following. As I said to you, you want to remove a small but known size of FTA paper and put that into a 96 PCR plate for downstream processing. For that purpose or for this purpose we have had this machine designed.

The basic premise of this machine is as follows: The robot arm would place down the tray containing the Marshal cassettes on the work surface of this punching work station, and those pneumatic cylinders over there - each of those have a collar which swings around and picks up a disposal plastic pin which is being fed from a bowl feeder in the back, which I'll show you in a subsequent picture. A bowl feeder feeds these disposable pins into a tube, which makes it possible for the pneumatic punch to pick up this pin, go around, swing around, position itself over the specific roll that needs to be punched, and goes down and punches out a piece of paper, which then falls into a PCR plate which is moving on a Cartesian Axis underneath the tray. By doing that we are capable of punching 96 samples into a PCR plate without any contamination or fears, without any issues as far as that is concerned.

Then the robot arm comes along and puts down the tray containing the Marshal cassettes. It then swings around and scans the tray to make sure that exactly the same device is being placed down. I hope I've got a picture of that because that's quite something to see. I don't think I've got this picture. What it then further does is it moves forward and it also scans the individual bar codes of the cassettes which it expects to receive through the LIMS as confirmation of which cassettes are in which position.

For the benefit of taking these photographs we had some of those glass panels of this device removed. From this point on forward it is physically impossible to effect even through malicious intent - to effect sample switching because after the robot arm has placed the tray on the work surface you cannot get manual access to the samples which are being punched down into the PCR plate, which is located over there and moves underneath the tray. This device is secured through doors which will physically bring the system to a standstill if you open them while the system is in operation. So it's impossible to effect sample switching even, as I said, if you really wanted to.

There you can see the bowl feeders responsible for feeding those disposable pins into the system. From this view you can physically see - given that's the robot track running over there, you can see how the robot has access to the punching work station from this side.

What I also wanted to show you in this slide here is this carousel, which is here. All of our magazines containing these 24 trays are loaded into carousels with turntables, which then positions the specific required magazine towards the robot track at any given time. We've got four of these carousels in the system looking after the loading of samples and the off-loading of samples.

As I said to you earlier on, when I speak about hitpicking, the important thing there is that we want the samples - if I say "samples," I'm talking about the reference DNA samples on the cassettes - we want those samples to remain in the system in the automated lab until such time that we are sure that we have received an acceptable DNA profile.

What happens then after the four samples have been punched from these four trays the tray is moved from the punching work station into one of these holding carousels over here, and eventually after we know that all the samples have given us DNA profiles that are acceptable, only then will the system release the tray to go out to be removed.

PCR setup again just takes place in one single liquid handler, and this device over here is just a cooling block which makes it possible for us to take our PCR master mixes and put it into the block let's say at 10 o'clock in the morning and leave it there until 10 o'clock the next day until we remove the empty tubes and replace it with full ones because the liquid handler takes care of mixing or doing the great mixing of master mix and then also dispensing the master mix into the PCR plate, which is situated next door.

As you can see over there, the previous one, the liquid handler is physically picking up some master mix from the cold store and dispensing it into the respective wells. Before we can get this plate sent through to the next lab we have to have this 96 PCR plate sealed with a metal foil, and that takes place on an automated laboratory plate sealer, which really seals those plates so well you physically have to destroy the plate to get the seal off.

So it makes it possible for us to transfer the samples without any fear of contamination, and also those of you who have crime lab experience and know what it takes to close down individual PCR tubes, this machine closes 96 or seals 96 wells in less than ten seconds, no sore thumbs, nothing.

This is just a view from the other side where the robot arm physically goes along and accepts the plate that has been sent through the transfer chamber. First of all, before it really accepts a plate it will go and check and see is this the plate that I expect to find or is this something that has been put in by someone else? So it goes along and it scans the bar code, which is situated on the PCR plate. It will then go and remove the plate from the air lock.

This air lock is a double door system where the one door will open, the door leading to the lab from where the sample is inserted. If I say "sample," the samples are inserted, the PCR plate is inserted. The PCR plate is then positioned on a turntable in the middle of this air lock, the door behind it closes, and it is physically replaced. In other words, we create a negative pressure in this air lock.

Then the air is normalized again, the door on the following side is opened, and the robot arm responsible for sample transfer in this lab, it goes along, recognizes the plate, accepts it, and picks it up and takes it into the lab for further dusting and processing.

DNA amplification takes place in the thermocycler. I'm sure you all familiar with those. In fact, when we designed the system, we were quite surprised to find that this was the only robotic amenable thermocycler available at the time that we knew of. We didn't know of anything else. So as I'll get to later on, all of these instruments are remotely controlled from a centralized control system, and I'll come to that just now.

What I wanted to tell you was we are looking at the thermocycler. The only two instruments that we could purchase commercially off the shelf that we could put down and integrate in a plug and play fashion were the automated lab plate sealer and thermocycler. Every single other machine had to be extensively modified to make it do what we wanted it to do.

I'm seeing Dr. Butler over here. This is quite an experience putting a face to a name.

Anyway, what we need to do is we need to have these 96 PCR plates pierced. We need to have those foil seals pierced before we send them to the liquid handler for transfer of amplified DNA samples and for the addition of internal standards and so forth.

So we have designed this monstrosity over here. It's pneumatically operated. What it does is the cylinder which you see running in the center over there, that cylinder goes along and it picks up a piercing plate. I don't have a picture showing you what this piercing plate looks like, but if you want to draw a picture in your own mind of what it looks like, imagine you've got a piece of aluminum with 96 small little pyramids on it. That's the piercing plate.

So what we do is we have this piercing plate picked up by the cylinder through an electromagnet. It then moves it along to where the robot places down the PCR plate containing amplified DNA, and it goes and pierces 96 holes into the foil. The plate is then sent out to be cleaned. The piercing plate is then sent out to be cleaned, and we've got 96 wells pierced or 96 holes pierced into a piece of foil.

This is the kind of thing that you really want to look at if you want to do this kind of thing. This machine, which is available commercially off the shelf, pierces those plates in an automated fashion, and you don't have to build this agricultural looking implement; you can really have a lab instrument. It works for us.

The central transfer for capillary electrophoresis setup takes place on another liquid handler. The robot arm goes along, puts down the pierced PCR plate containing amplified DNA. We've got another piece here called cold store built onto the work surface of this liquid handler, and the liquid handler then goes along and merely transfers samples from the PCR plate containing amplified DNA to the job plate, which is the one that needs to go into the 3100 genetic analyzer.

This was a challenge, getting the robot or the robotic system to effect this procedure which now follows. As you can see over there, what we did was we had this little clamp built onto the liquid handler which would hold down the plate that goes into the 3100 genetic analyzer, and what we do is we preinsert the rubber sytem, which is required for capillary electrophoresis downstream onto this plate so that the robotic device doesn't need to do that when we need it because, as Commissioner Morris said earlier on, our point of departure, our goal was to have no or minimal human intervention.

This system is designed to run 24 hours per day with a half an hour window period where you can go about changing your reagents, replenish consumables, and so forth. So we really couldn't have a person go in after everything is automated to go and build this little sandwich structure and put that into the genetic analyzer. That just defies the object of the exercise. I'll show you how we went about that.

Then again you can see this preinstalled on the plate for the capillary electrophoresis, and you see a picture of the robotic liquid handler effecting the transfer from the cold store of allelic ladders being put into the plate. What we do is we put allelic ladders in groups of six, one allelic ladder per two rows of samples just for practical purposes.

That is the automated storage cold store. Again, it's something that we had locally designed and built. It cost us about a sixth of what a commercial unit would cost us, and it has about double the capacity of one of the larger commercially available units.

The robot arm when it places a PCR plate from which amplified DNA has been transferred into the cold storage unit, the robot arm scans it just to make sure that it is placing exactly the right plate into cold storage. What I also just need to say at this point is we have another automated lab plate sealer in this lab, being the DNA amplification and detection lab, so that we can reseal the plate which has just been pierced, as I explained to you earlier on.

This same agricultural looking device, farm implement makes it possible for us to build the sandwich structure which is required by the AVI 3100 to effect capillary electrophoresis. This base plate thing - I don't know what it's actually called, but just to give you an indication of what this is, the black plate which you see over there at the bottom, that is a base plate, if you will. A 96 PCR plate then needs to be put into this base plate, fitting onto it, and then you have to go about putting a top cover plate over this whole assembly and clip it into the base plate so that you have a sandwich, for lack of a better description. So this machine builds that thing for us, no problem whatsoever. The base plate is being put on and it builds the sandwich for us.

Fragment analysis takes place in, as I said, the 3100 genetic analyzer, and for this purpose we also needed to modify an instrument which is sold as a fully automated genetic analyzer. Isn't that so? Now,the first thing we did was we took off one of the plastic covers from the doors to allow the robot arm to have access to the door. If I talk about the door, this structure here, this part is a door which you have to manually open. Then you have to manually put this sandwich device onto the work surface, you close the door, and you press the start button. We didn't want to do that, so the robot arm does that for us. We just took away that plastic part which was in the way, and the robot arm actually fits there and does what it needs to do.

It may sound like a very simple thing to get an instrument of this nature to be remotely activated to commence the start of a run. This was not so because this instrument is not designed, nor does it allow that by nature, and we had to work in conjunction with engineers from applied Biosystems to get this machine initialized by an external control system.

This is just another picture of the robot arm interacting with the genetic analyzer.

I'm just going to mention a couple of things about hitpicking in the last couple of minutes. Given the fact that we have these Marshal cassettes stored in a holding bay, if you will, and given the fact that we have amplified DNA stored in an automated storage and retrieval cold store, it makes it possible for the robot system upon input from an analyst who is now instead of using muscle power, using brain power, this analyst can say, listen, I want this sample to be rerun and I would like the rerun to commence from the start. In other words, let's go for the next well and punch the next well that comes through the system again or they can say we don't wasn't to waste all of this money that we spent on amplifying the DNA. Maybe it's just one of the capillaries that gave up the ghost, and we would just like to read that specific sample from the amplified product. So those actions can both be taken care of in an automated fashion.

The control system is quite unlike most other labs, most unlike anything that you may encounter in most other labs. This whole system is controlled by a programmable logic controller. We've got this one - we just call it mother. We've got this one mother PLC who takes care of all the motions that need to be fitted in the system. The human-machine interface to this PLC, which is a dumb animal sitting there in a cabinet which I'll show you a picture of - the human-machine interface is effective through a SCADA system, which is an acronym for systems control and data acquisition.

This is exactly what we want. We want to be able to sit in front of a PC screen and remotely adjust set points for temperature or the differential pressures in the lab. We want to get twins from that. So data acquisition on the one hand and systems control on the other hand makes it possible for a human to interface with the automated system, if you will, from a remote facility.

We run a tracking database on SQL server, which also links to LIMS and our equivalent of - I'm sure you're all familiar with True Allele. We've got an equivalent which was developed by one of my colleagues, Stewart Allen, in the STR lab.

This here is a picture of the PLC system and the devices underneath, which are all the server controllers looking after motion in the automated system. That gives you an idea of all the wiring that goes through the system. The engineers told me you can't take a picture of this because it's not neat. I said then fix it. They said to me we can't. So here is the picture nevertheless.

All of these wires come from the PLC. Just to give you an indication, you would have a PLC standing on the next floor, the automated lab also on the next floor, and then through the floor we had holes core drilled, and all the cabling would come through the floor and would then emerge through the core drilled holes underneath the instruments we required. So there is no breaking the integrity of the walls in the system. Everything comes through the floor, all the services that are required.

You may have noticed that there were some round disks in the lab. Those disks make it possible for the instruments to be swung around so that you can face the instrument when you want to service it or replenish chemicals, and then you press a button and the server moves the machine back to its logical position so that it can interface with the robot arm. This just shows the vacuum filtration unit to wash the FTA paper and the PCR setup machine over there facing away from the robot towards the human operator.

The throughput of the system is something that we were quite happy with. As I told you, we wanted the system to be run 24 hours per day with a half an hour window period during which operators would go in, one operator per lab, to replenish chemicals and consumables and so on.

This system given one genetic analyzer, which is very important - given one genetic analyzer, this system is capable of typing 440 samples per 24 hours. This is wrong. This one hour over here should have been up there next to the operators because it's one hour operator time per day. So if you would just excuse that.

So this automated system is capable of typing 440 samples per day, and given the fact that we don't expect any operator errors or anything like that, we then run the samples through the system once, which then gives us 440 DNA profiles. With the manual system currently in place we need six analysts to effect the processing of 840 samples, which works out to be 420 DNA profiles given the fact that we have to duplicate all the samples to check for errors and stuff like that.

So the automated system, per week we have two and a half hours of operator intervention, whereas in the manual we have 1,630 hours of operator intervention. Given the fact of the turn-around time of the 420 DNA profiles of five days, the manual system is capable of giving us 420 DNA profiles, whereas the automated system gives us 2,220 DNA profiles per five days.

What Commissioner Morris just reminded me of is we have left a space open on the work surface in the amplification detection area for an additional genetic analyzer, which is just going to double the throughput.

I'm not going to spend much time on this because I've got one minute left. These are just a couple of plans of the new facility that we're going to move into. It's an old abandoned hospital building. As you all know, when a DNA lab needs to be set up, they look for an old, run-down building, and say here you go and be thankful.

I'm not going to spend too much time on this. What I want to show you is just an overview. This gives an indication - if you will indulge me with more minute, this area over here will be our evidence recovery lab, and if you just look downstairs, that's the area over there. So what we're going to do is all of the crime sample processing for inclusion in the automated system will take place over here, and through pneumatic tubes we will then send those samples over to this wing over here, which will then be the automated lab setup.

Over here samples will be added to 48 well and 96 well lab plates, and through those devices over there, which are air locks, the samples will be introduced into the automated labs for processing. That over there, those are the tables on which instruments will rest, and that's just the robot arm that will effect all the motion in there.

So you see we've got one, two, three lab facilities lying adjacent to one another. What you see over there are antechambers because what we plan on doing is we plan on running these labs as if they were cleaning rooms. We don't have a clean environment, but just for cleanliness and good housekeeping we're going to do that.

Ladies and gentlemen, my time has run up one minute and ten seconds ago, so thank you very much for this opportunity. Commissioner Morris has told you how much we appreciate this, but just from a personal point of view this is really a fantastic and very generous opportunity that you have afforded us.

Thank you very much. I hope in this short span of time we've given you an indication of what we're doing, what we're busy doing, and we would love to speak with you over the next day and also tomorrow about what we're doing, and we would also be very interested to find out how you go about automating your labs, if you do that, and how you see things happening in the future.

Thank you very much for your time. Thank you.