Cell types

[shruti...@gmail.com]

Hi!

Is there going to be an addition of some of the more common cell types in human, soon?

Thanks!

[michael]

Hi,

methylCRF is premised on the hypothesis that our training method allows a model trained one cell type to be general enough to accurately predict methylation in any cell type.   In the paper, we showed that using our H1-trained model showed equivalent accuracy in non-ES cells.  Since then, we've even used this model on several others species with similar results.  So, please feel free to use the H1 model on your data.  Please don't hesitate to ask if you have any concerns or questions.

best,
michael

[bioin...@gmail.com]

Hi Michael,

How distant are the other species that you have tested? I'm new to this area and am interested in applying methylCRF to some insect data. Is it possible to generate insect-specific data files and MRE fragment files or should I go back to the method outlined in Harris (2010).

kind regards,

Kate

[michael]

Hi Kate,

Thanks for the question.  The furthest we've tried is zebrafish, and we've gotten good results.  Since we use annotations of TEs, CpG islands and such, my guess would be that insects might be too distant to human.  A quicker option could be to use another method from our lab that estimates differentially methylated windows between to samples without the use of assumptions about a genome, M&M:
http://genome.cshlp.org/content/early/2013/06/26/gr.156539.113.abstract

If you are interested, you can email Bo or I can have him email you.  Otherwise, if you are interested in methylCRF specifically, there is however, a bit of work involved with porting it to each insect species.

best,
michael

[bioin...@gmail.com]

Hi Michael,

Thanks for the clarification. I will email Bo about the MnM algorithm. As I am looking for hemimethylation in one sample and not a difference in methylation between samples I did not think it could be applied - but he may be able to advise me about it.

Kind regards,

Kate

[arneb...@gmail.com]

Hi Michael,

I would also be interested in the feasability of porting methylCRF to an insect species. How much effort do you think this would be?

Best,

Arne

[bioinformatics girl]

Hi Michael,

I don't seem to have got through to Bo about the MnM algorithm - I may have the wrong address. Could you please forward my email?

Kind regards,

Kate