Things to do:
- Reference all imputation notes from NCBP manuscript. Apply method 7_1 to all data from PC and DC. We expect to have more replicates for DC by the time of re-submission.
- Find a strategy to deal with isoforms. John suggested that in cases where there are enough unique peptides to identify a single isoform, indicate which one it is. Otherwise, indicate only the protein and in a separate table, put the isoforms identified and which are the most likely.
- Compare our data with the data from previous experiments. Compare the PC and DC data with the data from I-DIRT (from Taylor 2013 and Taylor 2018), the CRC paper and Vuong 2019.
- Highlight the proteins previously detected in the PC and DC plots for each of the above studies - using a different color to mark the factors of each study may be interesting.
- It would be interesting to highlight which factors have been found only in PA-1 PC and Vuong 2019 and are not in the DC, I-DIRT and CRC data. - GO analysis. Repeat the GO analysis using only the unique factors of mono and nano. Try using a more stringent threshold for the p.adjusted value. Use only the data from method 7_1.
- Try to create a MDS plot instead of a PCA.
- L1TD1 (ECAT11): It is interesting to find this factor in our data, however this paper (at the moment) is not the place to talk about it. Save this data for the future. It seems that it has been captured only by LaORF1, we know that this factor has perinuclear expression, apparently associated with the Golgi.
- This suggests that the nanobody may be capturing ORF1 present at this cellular location.
- Note: since L1TD1 is so similar to ORF1, the nanobody could be directly detecting ORF1?
- However, we know that ORF1 and L1TD1 co-localize.
- John suggests a possible inhibition of LINE-1 by competition with L1TD1.
- Mehrnoosh thinks that there is evidence that L1TD1 might bind to telomerase like factors to play a similar role to ORF1 for L1. She says she found evidence that L1TD1 may interact with BCLAF1 and TRAP3.
John notes from summary email:
1. We SHOULD attempt to capitalize on differences between PC and DC IPs that support interactome differences rationalized through the different cell biology of these cell types - this *may* be visible through certain proteins that are differentially co-purified and/or through GO enrichments
2. We SHOULD NOT attempt to use the data as a way of showing that these cell types are different (not sure if I understood this point) - this has already been done by other methods. We cannot use these IP data in the same way you would use a “whole transcriptome” data set - since all our observations are anchored on ORF1p - it could be ENTIRELY POSSIBLE (although I admit, very unlikely) for the ORF1p interactomes to be different in PC and DC even when only a small % of the proteome changed at the macroscopic level. However, it is well now that if you take any two cells, ~50% of the proteome is different (rule of thumb) - so, gross proteomic differences are not usually very informative without other pieces of information. I therefore expect PC and DC proteomes to be different but without doing the experiment I cannot say how this correlates with the IP differences we see.
3. Therefore, we could choose to do the whole proteome experiments IF we think they will contribute to the paper - and we could not need much protein for that - so, maybe we can do that sooner than later as well. We could then see how our interactome differences line up against microscope proteome changes. I am not against this experiment.
4. We agreed that the L1TD1 story is for another manuscript - but we may find reasons to *hint* about our next manuscript in the discussion section of this one - L1TD1 should not be a focus
The slides from this meeting are available in ppt format here