Hello, has anyone had experience with optimizing easyBlood image processing. Our lab has spent a lot of time trying to calibrate and optimize the image processing and liquid handling in the Hamilton easyBlood system. But unfortunately in larger tube runs or with poor samples we get stuck in loops of 7 out 8 tubes are perfect but 1 just can never find the buffy coat.
If anyone has worked with these in the past or has any tips and tricks, it would be much appreciated!
Given the variability of blood samples, there could be a number of reasons that some samples aren’t cooperating, so providing optimization tips over the forum may not be the most effective approach. For this situation, local, on-site support from a field application specialist with eyes on the samples and imaging software will be the most efficient route to a solution. Once resolved, we can always follow up here with a resolution.
If you need any assistance being pointed in the right direction, please message me directly and I would be happy to assist!
Thanks.
-Nick
Nick, Appreciate the quick response. Definitely agree on the points you made regarding the optimization and liquid handling.
I guess my main questions was if anyone has played around with manual entry of buffy coat layer height. For those samples that are so bad no amount of imaging is going to work. Our workflow often gets held up on those one off samples while the rest of the lot is totally fine.
Scientists have requested a sort of “manual override” to just get the method to continue on. I traced out a bunch of stuff and it looks like those heights are passed back into the easyBlood software for harvesting.
Of course improving our imaging and handling as well as error handling is always a good route too.
Hi @GrayHorn,
Which version of easyBlood are you running?
easyBlood 3.3.1 has a feature for manual fraction confirmation/manual adjustment. It’s a boolean value (i_blnManualFractionsMode), so you either get to confirm/adjust every single picture taken in each function call (generally speaking, one carrier’s worth of pictures), or you leave it to the software as you currently have it being done and not have an option to adjust manually. There is no option to split it up and it’s not designed to operate as an “error recovery” behavior where it only presents the manual adjustment on a failed image, it will do it for every analysis performed using easyStep_FractionLevel_Analyse_ManualFractions.
Sincerely,
William
Hey GrayHorn,
I had some sucess increasing the size of buffy coat in the past. All the information are coming from your image analysis software. That is were you could focus your effort. I dont the software any more, but they are settings for determine where to start looking for the buffy coat. I widened the searched band a bit and was able to triple the size of the buffy coat pellet. You are getting some blood and plasma, but those should come off during centrifugation if i recall. Its been a few year, but i would suggest optimizing the image analysis. The code for the easyblood are pretty solid.
Good luck!
Thank you this seems like exactly what we need. Will investigate, I believe we are using 3.1.0 though so unsure if that supports it?
Hello @GrayHorn,
The feature was implemented in v3.3.1, so v3.1.0 won’t have it. You would need to upgrade to v3.3.1, which may require some rewriting of your Method if any customizations were added from the demo Method that comes with the software install. Otherwise, if the base demo is what you used in the first place, the new demo it comes with would be a mostly direct replacement. But if you need to adapt existing Methods for v3.3.1, there will be some effort involved, as new functions were added and old functions had additional parameters added/changed between those versions, so your current function calls would be incorrect in the easyBlood SubMethod Libraries.
I would highly recommend reaching out to your local Applications resources, both for the software (we cannot share it in the forum due to licensing of the imaging software component) install and for any changes that may need to be done on the Method level.
Sincerely,
William
Thank you so much for all your help William. Will definitely reach out!