On Jun 29, 2012 4:58 AM, "Máté Ravasz" <ravaszmeister@gmail.com> wrote:
>
> Hi,
>
> For example I would love to have a reliable colony counter program.
>
> The job there is to take an image, like this: http://x53.xanga.com/341f440279533263478321/b210037979.jpg and count how many spots are there on each plate. The program needs to count the spots, but needs to exclude air bubbles and other artifacts.
>
> There are a few programs like this out there, but the ones I've tried give a lot of false positives and don't count the colonies very accurately. It can happen however that the perfect program exists, I just haven't tested it yet...
>
You mean not just a dumb tap the screen to count a colony? I.e. machine vision? There was a crappy Asian app on the android store that did that, you helped it along with false positive or negatives but I don't think it learned via the user input (you weren't training it). Something like this would be easy to implement with opencv or maybe one of the libs based on it, not sure how to accomplish the training though.
>
> Secondly, I would be very grateful for a program that can identify homologous proteins in different organisms. The idea would be to give it either a DNA sequence, or a protein sequence and an organism. The program would then find other organisms which have the same gene, or a highly similar gene in them. So far there are numerous programs that can do this. But the problem is that distant relatives are not found after a while, as these programs always try to match new hits to the original sequence. So I would need a program that not only searches for homologs, but after it finds one, it matches new hits also to the newly found sequence. This way even distant relatives can be mapped. I do not know of such a program, but it can happen again, that this has been done by someone before...
>
Wouldn't a program like that eventually find ALL or MOST other genes in the database depending on how loosely or tightly you define homologous?
--You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diybio@googlegroups.com.
To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/diybio?hl=en.
0 comments:
Post a Comment