Add human and mouse light chain models

[Jason Vander Heiden]

From Ang:

Steve said I could bug you to help me integrate the models to the website/packages in a user-friendly manner. Could you please help me with the following tasks?

1. Susanna brought up the issue of naming convention. Here's how I named the models: 1) hS5F-H 2) hS5F-L 3) mRS5NF-L. What do you guys think? These names are in the @name field in the R objects attached.
2. Can you please confirm that these 3 models have all been integrated into shazam? I have re-attached them in this email.
3. Would it be possible to add the targeting model RData and text files somewhere on the website? Right now users would need to install the entire pipeline to be able to access these models. I'd imagine that some users may want to just look at the values in a text file. Preferably on this page because this is where our paper points to.
   • http://shazam.readthedocs.io/

4. Can you please add citations for the mouse model to the following pages:

   • http://shazam.readthedocs.io/
   • http://shazam.readthedocs.io/en/latest/topics/MRS5NFModel/
   • http://shazam.readthedocs.io/en/latest/topics/shazam/

   Cui A, Di Niro R, Vander Heiden J, Briggs A, Adams K, Gilbert T, O'Connor K, Vigneault F, Shlomchik M and Kleinstein S (2016). A Model of Somatic Hypermutation Targeting in Mice Based on High-Throughput Ig Sequencing Data. The Journal of Immunology,197(9), 3566–3574. http://doi.org/10.4049/jimmunol.1502263

5. Can you please add citations to the shazam package as well?

Comments:

• Do not replace the human heavy chain model, and don't add the new heavy chain model. The old model is fine.
• Add the human light chain model.
• A mouse model is already in shazam, but we should verify they match and make it explicit that it's a light chain model.
• We should discuss how to redo the naming, and be as explicit as necessary, by adding something like IGH, IGL, IGK or VH, VL to the names.
• (3) is possible, but I don't think we should do it, as it'll be messy and inconsistent. Installing from CRAN is simple.

[Namita Gupta]

I think (3) should be added to clip.med.yale.edu/shm, which is where the older models were made available independently from the shazam package. Hailong can help with adding things to the website.

[Julian Zhou]

• Previous version of MRS5NF (also from Ang and added in the summer) differs slightly (mostly in the slot @substitution) from the latest file from Ang. New MK_RS5NF is based on the latest file.
• Added HH_S1F, HKL_S1F, MK_RS1NF as normalized 1-mer substitution models (actual rates, not distance). HKL_S1F is from Table 3 of Cui et al. (2016). MK_RS1NF is not reported in Cui et al. (2016) but was obtained directly from Ang.
• HS1FDistance and M1NDistance are kept unchanged since no conclusion has been reached as to how their names should change.

[Julian Zhou]

3 & 4 -> new issue