Calibration

Well, in short Calibration is hell.

Until we perform a calibration test on the AP96, and the Span-8 we do not know what volumes are being moved around. After two days of failed efforts with Fluorescein dye I reluctantly switched to a more simple, but less fancy, gravimetric method of volume estimation.

In the end I got the following results:





the second method name is gravimetric calibration june 27 07 ii, the first is without the "ii"

It is important to consider pipetting techniques when doing a calibration. The first gravimetric test I performed prior with different settings yielded different results. One must also consider evaporation as a significant source of error.


---
Tips to get the volume you desire

Let's start talking about methods where the aspiration is from a reservoir. It is important to aspirate at the proper height. When dispensing, the dispense velocity must be appropriate to expel the fluid without leaving a remaining amount in the tip or on the tip surface. A tip touch may or may not be useful in these instance. If no tip touch is allowed you can increase your blowout volume, but your largest aspiration volume will limit this. Any blowout around 50ul is considerable, but depending on the situation you may want to increase or decrease it.

More tips later.

Also you'll get a different behavior depending on fluid type.

Updates and lessons learned

Sofware upgrade story

Well lots has gone on with the robot recently.

To make a long story short the tech support guy, Phil, suggested to update the Biomek software to build 14. He couldn't do it since he didn't have the software on him, but had the software sent to me.

I did it on my own, since he was on vacation, and it was a learning experience. I called into phone support for guidance, and here is how it went.

So it is important to know that the "export all" funciton under the file option does not really mean "export all". This only exports methods. So they should call it "export methods" rather than "export all". So if you are doing a re-install remember to do a separate export for methods, instrument files, decks, hardware, etc. Otherwise you have to start from scratch.

I got lucky and found some files to recover, but it took me a whole day to rebound from the "export all" mishap.

Also, since Norton's Ghost is on the Biomek computer I suggest that you clone the whole drive on DVD and recover in the event of disaster.

---
Other news--Probe Replacement

Probe 7 of the span-8 was leaking ever so little, so Phil sent a replacement part and I replaced the whole probe. The leaking has gone away. If you experience any leaking be sure to call tech support. The liquid is from the inside of the machine and will contaminate your sample as well as cause a pipetting error.

---
Dilution of powdered chemicals and Tranfer from file feature.

The Cutler lab has a new library, but it came in powders. The library requires dilution with DMSO. Lots of lessons were learned about the tranfer from file feature.

Each chemical has a unique molecular weight, and as such requires a unique volume of DMSO to achieve a constant molarity. This task requires tranfer from file. It took me about four days to figure out how to do this, and with some intervention from Beckman forum users I discovered that if one wants to use a reservoir as a source they must call all the wells "1".

Don't bother looking in the manual, as it does not even support a source reservoir.

Suped-up Biomek

Ok,

It's time to get Biomek to do interesting and complex things.

Goals:

1. Get Biomek to handle Filled chemical plates from the hotel

2. Have Biomek dilute a powdered chemical library using a variable filling function from an input csv (excel sheet)

3. Use liquid level sensing to "discover" the liquid level and therefore the top of the liquid to accomplish variable aspiration from plates that have been cherry picked or have variable volume.

4. Combinatorial mixing. Taking two compound plates and mixing them in a combinatorial fashion to yield a plate that is a mix (synthesis) of two or more compounds.

5. Use barcode reader to inform the method what the incoming labware is and how to handle it. Use this function to have each library plate dealt with accordingly.

6. Remote Viewing. I would like to set up the Biomek computer on the network to be accessable via Windows remote desktop. With this capability the progress of an automation run can be periodically checked from a user's remote computer. In short you don't have to run to the Biomek room to see if the robot has screwed up or has stalled.



Solutions in progress:

1. Slow down the belt speed to prevent splashing, also define a new labware type and restrict handling to a low speed (5% or lower)

2. Use the Quant/Normalization wizard.

3. Perhaps some mix of the qant/normalization wizard with LLS output.

4. May be as simple as a generated csv input

5. Dunno, will work on it with Phil

6. Get CPU on network and install windows remote desktop.