PDA

View Full Version : Agilent Bioanalyzer refuses to size bands


pmiguel
12-09-2009, 11:45 AM
A perennial irritation with using the Agilent Bioanalyzer for DNA or RNA QC is that it frequently refuses to convert the x-axis from time (seconds) to size (bases). In our most recent case the issue that seems to trigger this stubbornness on the part of the application is that the data was collected for 69.8 seconds instead of the 84 seconds the application believes it should have collected data for.

I know of no way to control this behavior. Any ideas?

I've been through all the trouble shooting sections in the manual, but none of them address this issue.
--
Phillip

Rudi_Ratlos
12-10-2009, 12:03 AM
Hi,

the sizing does not work if
- marker(s) were not recognized correctly
- ladder wasn't recognized correctly.

if you can't solve the problem by reassigning markers and adjusting peak find settings, I suggest to contact Agilent on-line tech support at

Bioanalyzer Line: (800) 227-9770 x3 x4 x1 Direct Line: (800) 894-1304 x5454
or
bioanalyzer@agilent.com

Cheers
RR

pmiguel
12-10-2009, 07:19 AM
Hi,

the sizing does not work if
- marker(s) were not recognized correctly
- ladder wasn't recognized correctly.

if you can't solve the problem by reassigning markers and adjusting peak find settings, I suggest to contact Agilent on-line tech support at

Bioanalyzer Line: (800) 227-9770 x3 x4 x1 Direct Line: (800) 894-1304 x5454
or
bioanalyzer@agilent.com

Cheers
RR

Hi Rudi,
Thanks for the answer. I may contact Agilent as you recommend. But I do not know anyone who has used a Bioanalyzer frequently who does not run into this issue on a regular basis. So I think it worth discussing here.

I have used the methods you mention previously to solve some "failure to size" issues in the past. But there is another class of these that is particularly maddening. That is, there is no failure to size per se. Instead the program refuses to convert the x-axis to size (bases). That is, using other methods (eg "smear analysis") the program will correctly label the sizes of peaks. Therefore the program has recognized the ladder and and the markers were recognized correctly. Nevertheless, it proves impossible to force the program to convert the x-axis from seconds to bases.

This phenomenon seems particularly prevalent on pico RNA labchip runs.

pmiguel
12-10-2009, 08:27 AM
It seems that any errors (which appear as red dots at the top of a lane) in any lane will prevent the x-axis from converting to size (nucleotides). This would just be a plain old bug in the code. Either that or a design flaw in the work flow. There is no sense to it.

The workaround is to clear those errors by any means necessary. Unfortunately this is not necessarily trivial. On the pico RNA lab chips it is not unusual for the lower size marker not to appear on the electropherogram for one or two samples. If the software has called any peaks for any reason, you can right-click on one them and one of the choices is to set that peak to the lower size marker. This should clear the error. Once all errors for the run are clear, the x-axis can be switched to size (nucleotides).

If the software has not called a peak, then it is much more difficult in RNA assays. Beyond the scope of this post. But, in brief, it seems to require adding a fragment analysis to the "local settings" pane and setting the integrator parameters to a low enough stringency to call a peak in the interval set by the fragment analysis. One hurdle here is that even though the time<->size button will not change the x-axis, maddeningly, it does switch the unit in the "add fragment" table. So, if you happen to have it set to "size" and don't notice, the interval you set will be invalid ("nt" instead of "s") and you will never be able to convince the software to call a peak.

--
Phillip

Roger Adnan
08-25-2010, 01:04 PM
0) as noted, this is a real pain in the software, and it is one of many pains in the software. If you don't complain to agilent, or tell them that you are not thinking of the new shimadzu (reusable cartridge) or the biorad, they won't do anything - btw, the least they can do is send you a free kit; if enough people complain, maybe they will do something.
Remember, the 1st law of business is, if it isn't impacting sales, it isn't a problem.

1) the workaround that I found was to set the "peak height" value (under global settings) to some very low value - when you do this, the software will call very small (noise) peaks as peaks, this allows you to set an upper and lower marker in each lane, which at least allows the software to work.

2) there is also a problem in that the x axis values are not "real" - if you assign some peak to be the upper or lower marker, note that the x axis value, IN TIME, changes - this means that that time scale is some sort of computed scale, not a real scale
3) Another big problem with the software is that they don't give you an equation giving bp as a function of time, which would be nice if you wanted to export data and do some stuff
4) the results reporting is not very useful, I find myself having to do huge amounts of work
5) you can't set the region to a peak

pmiguel
08-26-2010, 04:30 AM
Hi Roger,
0) In general, I agree. But many of us are used to the support we get on instruments that cost half a million dollars and for which we pay fifty thousand a year in service contracts. Since the Bioanalyzer costs "only" $20K or so and the software glitches are an occasional issue, there is likely to be little complaint.
1) Yes, that is one I use also.
2) I'm sure that is just an offset, right?
4) Seems like most people just accept what the software gives them. For example, the completely fabricated metric the "RIN" score, for total RNA quality, is accepted without question by most. But only the most vague description of how the software comes by the "RIN" score is given.
5) And the tricks one can use to accomplish this indirectly (eg "smear analysis") are different for different analysis (chip) types.
--
Phillip

Roger Adnan
08-26-2010, 06:24 AM
Normally, I would agree with you about the cost thing - you can't expect the level of service you get on a half million dollar HPLC/MS with a cheap instrument

However, the bio analyzer is a bit unusual for the lab market, in that the volume is very, very high (for the lab market). Thus, we can expect better.

More generally, as an academic turned industry guy, I am constatnly amazed at how accepting scientists are of obvious flaws,
Agilent is making a ot of money on this instrument; we have the right to demand better software.

NextGenSeq
08-26-2010, 07:09 AM
Agilent has always had very poor technical support. I'll never forget when Silicon Genetics was bought out by them and then they moved all the tech support to India and hired people who couldn't speak English well.

Caliper sells their own new instrument now (they developed the Bioanalyzer for Agilent). Hopefully their support is better.

Susanne
08-27-2010, 04:57 AM
There's a little info inside the software on the RIN - go to Help - Contents and Index. in the Application Notes under "Related Documents", there's something on RIN, and also you could search for RIN.

Susanne
08-27-2010, 05:06 AM
Referring to the time shown (x-axis) that changes when the markers are changed (DNA chips): in order to calculate the correct sizes for all lanes, a 'virtual time' is assigned to the markers, and these are aligned with each other (all lanes) and with the Ladder. So if you change a upper or lower marker, clearly, the time annotation of the x-axis does change. To see the "real" migration time, just turn off the analysis (stop button).

VanessaS
08-27-2010, 10:32 AM
Has anyone tried the Caliper instrument?

SABC
03-08-2012, 08:34 PM
There are instances where this solution will not work, for instance if the run is stopped before the trace data begins for a particular sample. The result file is just an empty xy graph, with no chance of assigning markers.

I tried using the Comparison context to import the working samples, leaving out the failed sample, but even with no failed samples present in the comparison window the x axis wouldn't switch from time to size.

Susanne
03-09-2012, 12:32 AM
Unfortunately, in the comparison context, it doesn't work to change the x-axis to size. Only time is available. :(

pmiguel
03-09-2012, 04:29 AM
I don't use the Comparison context much. But I did not think it ever displayed size -- only time.

One last possibility I can think of in your case:

Click on the "Files" tab and choose "Save selected sample...". Choose only the lanes that did complete. Save the new, partial .xad file. Open that file. See if it will now size your samples.

--
Phillip