Is it possible to detect 5hmC together with 5mC? Also, in non-CpG content (for 5mC). Is there an update to this video coming? This video is still linked to PacBio's website when looking into epigenetic detection with PacBio.
Hi Aaron, Does the training data have to be from the same genome? In other words, can you train the algorithm using a human data training set and then apply it to a different organism's (non-human) genome?
Hello, I am not Aaron but yes and no. If you train a model from a human genome and then apply it to mouse data, things should go pretty well. If for example you try to apply it to an insect or even a plant, where the genomes and epigenomes substantially differ, this would not be suitable. So, in principle you would want to use a closely related organism, but that of course also depends on your research question.
Is it possible to detect 5hmC together with 5mC? Also, in non-CpG content (for 5mC). Is there an update to this video coming? This video is still linked to PacBio's website when looking into epigenetic detection with PacBio.
Would be very cool to have asymmetrical modified base detection. mCH is prominent in neurons and this method wont be able to capture that.
Symmetry is a design choice of our current algorithm, but it could be extended in the future to consider strands separately. Thanks for the comment!
Hi Aaron, Does the training data have to be from the same genome? In other words, can you train the algorithm using a human data training set and then apply it to a different organism's (non-human) genome?
Hello, I am not Aaron but yes and no. If you train a model from a human genome and then apply it to mouse data, things should go pretty well. If for example you try to apply it to an insect or even a plant, where the genomes and epigenomes substantially differ, this would not be suitable. So, in principle you would want to use a closely related organism, but that of course also depends on your research question.