post biodesign phd oral defense 04-12-2014d1734

2014-08-29

post biodesign phd oral defense 04-12-2014d1734

defense was held 4-11-14

pass/fail form partially filled
"C:\Users\kurtw_000\Documents\kurt\storage\CIM Research Folder\DR\2014\04-21-2014d2130\Kurt Whittemore Report for Dissertation Defense.pdf"
pass/fail form fully filled 04-21-2014d2130
"C:\Users\kurtw_000\Documents\kurt\storage\CIM Research Folder\DR\2014\04-21-2014d2130\4-21-14\Kurt Whittemore Report for Dissertation Defense.pdf"

audio recording of presentation
"C:\Users\kurtw_000\Documents\kurt\storage\DR\Audio 2014\04-12-2014d1446\PhD oral defense 14-04-11-15-24-45.wav"

Video recording of phd oral defense



notes from audio recording 


questions after presentation start at 1 01 50





Diego


you didn't check if there was an immune response against the antigens you found





me


the screen was testing an immune response but I didn't use alternative methods





Diego


why would entropy change as people get older





me


affinities and antibody repertoire changes with age





Diego


difference between complexity and randomness.  careful with terms





Neal


a completely non random thesis world be bad if it were all as (letter as).  order of letters matters.  what if all the binding events on the array are not independent.  how could you determine interconnectedness of binding events





me


statistician zipf and power law distribution of letters and words. . .maybe complex mathematics could find interdependence of features on array





Bradley


asked about lab dreams project animation.





me


progressing slowly





Krishna


were the identified tumor antigens previously associated with cancer





me


yes with the myc mad max and others





Krishna


did you look at function.  were they house keeping genes





Me


yes.  also house keeping genes are related to cancer in other studies





sai


asked if bacteria are the best for making mammalian proteins





me


may not always be the best but most practical





public audience questions ended at 1 11 08





exam part with committee starts at 1 19 08





Phil


what further experiments could you do to relate entropy measure to health





I talked about sera from athletes and non athletes





Phil


how would you decide number of groups and class sizes





me


standard deviation, power, the test





Phil


what about infectious and chronic disease groups





me


expect low entropy for infection and high for chronic





valiery


better explanation of idea behind this


question about lymphoma and vaccine





me


both lots of antibodies against few targets





Stephen


you could have followed up and looked up numbers for antibodies in blood after vaccine or infection





Kathy


comment about be cell cancers





me


dog lsa is a b cell cancer





Bert


interested in normal and cancer mouse.  what was the system strain etc





me


fvbn neu t mice tumor at 33 weeks.  thus one had tumor at 33.1 weeks.  could detect peptides increasing 8 weeks before.  maybe immune system battling tumor





discussion about response to her2





Stephen


give therapy at peak





immunosignature doesn't look chaotic





Bert


entropy of entropy (over time)





Stephen


tiger has more samples





Bert


trying to understand monoclonal entropy slide





Kathy


comment about avidity





Stephen


he thinks avidity is independent of specific binding.  many high affinity antibodies have a more rigid structure which could help explain result.when they bind they bind tight and don't bind much else.  previous experiment in our lab showed that antibody diluted into sera bound just as well





Stephen


practical idea for work.  quantify antibody specificity for antibody companies.  people are worried about cross reactivity.  I thought this was a really good idea.





Phil


that would be cheaper than proto arrays





Bert


idea of looking at hiv people and comparing those that develop broadly neutralizing antibodies





Phil


if you were given a lot of money how would you redo the experiment.  he asked what was the hardest part





me


I answered wrong by saying I would screen a lot more (committee didn't like this answer).  then I said I would screen the spot with highest p value which was decreasing with tumor sera





Stephen


why didn't you have shen and Felicia try to vaccinate mice with the clones you found?





me


everyone was so busy





Kathy


is it necessary to include other measures other than entropy





me


yes I got better performance in my own tests





Kathy


maybe include goodness of fit of one distribution to the other. . . and peptide sequence information.  counts of all peptides with a similar sequence. 





Kathy


maybe some of the measures could be thrown out





Kathy


how did you rank entropy





me


four different experiments and p value between disease and normal





Kathy


did you select certain features before calculating abstat





me


no





Stephen


why not just do ages with scatter plot





me


I did but it didn't work well





they were interested in the r2 of the scatter plot as well as the ages.  I told them the old ones were all in 50s and 60s





Stephen


what are requirements for system for monitoring





me


needs to be cheap





Stephen


joke.  this is so useful.  you can tell people how old they are





exam section ends at 2 01 39





Stephen





rhinopest has also been eradicated





if you have a pointer just point don't wave





didn't make it add clear that I was taking about preventative vaccines





didn't explain that I needed sera from someone with a tumor and someone without


presentation too fast





Bart thought entropy explanation was clear





Kathy


you can make an exception for champagne just once





thesis so imaginative and unique with new and interesting ideas.  I felt like I'm an art student haha





Stephen


you didn't finish everything you started





got laser pointer as a gift from cim