Friday, October 18, 2013

Journal club: Ontologies to hypotheses. UhLAWLNo

The design for data generation and the actual data generated seem useful and adds to the field of characterizing totipotent cells. The authors' build bridges from gene ontologies of hierarchical clusters to hypotheses with little to no support in between, for example when they found that ribosomal protein genes were enriched in four clusters, implying differential expression, they hypothesize that "ribosomal equipment needs to be precisely regulated during re-entry in to the cell cycle". But, what if this is just peripheral information that is non-essential to the gene state, and that even when it is not precisely regulated the cell can still enter the cell cycle. The analyses seems peppered with weakly supported hypotheses.

I think the authors' story was poised to take better advantage of the many similar data sets generated by other labs. And, I think that had they made better use of the other datasets, they could have
  1. potentially given their hypotheses more support, and 
  2. I think more importantly, make a stronger case for their dataset's contribution to the field.
During journal club, another student asked why they didn't collect meristem cells if that is essentially the state they are trying to assay; and it dawned on me that these protoplasts may be an "erased" state, and have no tissue identity, not even as quiescent cells. This leads back to Dr. Maggert's question, with respect to epigenetics, if there is a temporal/spatial difference between "erasure" and "establishment", and I think that this could be an example of a cell that has been "erased", but not yet "established". I think that Dr. Maggert would 100% argue that this event falls into the category of development, which he defines as mutually exclusive to epigenetic. 

I appreciate Dr. Devarrene's insights in Journal club. I think they add value to learning how to critique science.

No comments:

Post a Comment