A Conversation with Robert C. Elston
aa r X i v : . [ s t a t . O T ] J u l Statistical Science (cid:13)
Institute of Mathematical Statistics, 2015
A Conversation with Robert C. Elston
Gang Zheng, Zhaohai Li and Nancy L. Geller
Abstract.
Robert C. Elston was born on February 4, 1932, in London,England. He went to Cambridge University to study natural sciencefrom 1952–1956 and obtained B.A., M.A. and Diploma in Agriculture(Dip Ag). He came to the US at age 24 to study animal breeding atCornell University and received his Ph.D. in 1959. From 1959–1960, hewas a post-doctoral fellow in biostatistics at University of North Car-olina (UNC), Chapel Hill, where he studied mathematical statistics.He then rose through the academic ranks in the department of bio-statistics at UNC, becoming a full professor in 1969. From 1979–1995,he was a professor and head of the Department of Biometry and Ge-netics at Louisiana State University Medical Center in New Orleans. In1995, he moved to Case Western Reserve University where he is a pro-fessor of epidemiology and biostatistics and served as chairman from2008 to 2014. Between 1966 and 2013, he directed 42 Ph.D. studentsand mentored over 40 post-doctoral fellows. If one regards him as afounder of a pedigree in research in genetic epidemiology, it was esti-mated in 2007 that there were more than 500 progeny. Among his manyhonors are a NIH Research Career Development Award (1966–1976),the Leadership Award from International Society of Human Genetics(1995), William Allan Award from American Society of Human Genet-ics (1996), NIH MERIT Award (1998) and the Marvin Zelen LeadershipAward, Harvard University (2004). He is a Fellow of the American Sta-tistical Association and the Institute of Mathematical Statistics as wellas a Fellow of the Ohio Academy of Science. A leader in research ingenetic epidemiology for over 40 years, he has published over 600 re-search articles in biostatistics, genetic epidemiology and applications.He has also coauthored and edited 9 books in biostatistics, populationgenetics and methods for the analysis of genetic data.The original conversation took place on August 4, 2009, during theJoint Statistical Meetings (JSM) in Washington, DC by GZ and ZL.NLG had dinner with RCE during the 2013 JSM in Montreal, Canada,and added supplementary material and edited the conversation. RCEupdated and clarified certain points.
Key words and phrases:
British statisticians, family studies, geneticepidemiology, history of statistical genetics, pedigree data, S.A.G.E.software, statistical genetics, statistics biographies.
Gang Zheng was a Mathematical Statistician, Office ofBiostatistics Research, National Heart, Lung and BloodInstitute, Bethesda, Maryland 20892-7913, USA. Hepassed away on January 9, 2014, without completingthis. Zhaohai Li is a Professor of Statistics and Biostatistics and the Chair, Department of Statistics,George Washington University, Rome Hall, 5th Floor,801 22nd St. NW, Washington, DC 20052, USA e-mail:[email protected]. Nancy L. Geller is the Director of Office ofBiostatistics Research, National Heart, Lung and Blood G. ZHENG, Z. LI AND N. L. GELLER
1. EARLY EDUCATION
Gang and Zhaohai:
Robert, it is a great pleasureto have this opportunity to talk with you about yourlife, research, career, mentorship and some of yourviews of genetic epidemiology.Can you begin by telling us about your earlyyears?
Robert:
I was born in London, and I was 7 yearsold when World War II broke out (1939). My broth-ers and I were evacuated to a little village, LeaGreen in Hertfordshire, about 30 miles from Lon-don. That’s where I first loved farming and thoughtI’d be a farmer when I grew up. In 1941, my fatherarranged for us to live in Hertford, where BatterseaGrammar school had been evacuated from London. Idon’t know how he got me into that grammar schoolsince I was really too young. They put me in the low-est form (grade). I eventually took what was calledthe school certificate at 14 while most took it whenthey were at 16.
Zhaohai:
What did you study in high school?
Robert:
In addition to the usual subjects, we stud-ied French and a year later got to choose Latin orGerman. I did Latin mainly because my elder broth-ers had done Latin, and because I knew I neededLatin to go to Oxford or Cambridge. The followingyear, the class master who taught Latin chose thetwo or three best students and said: Okay, you willdo Greek. My brothers had done Greek but theyhad to give up physics to do Greek and I was notgoing to give up physics! I said, if you want me to doGreek, I’ll need to eliminate history or geography orboth. So they agreed!When World War II ended, Battersea Grammarschool moved back to London, so I had one year atHertford Grammar school. In 1946, we returned toLondon as a family. Then I went to University Col-lege School, which was “a public school,” meaning itwas open to anyone who was willing to pay (laughs).Although I studied Latin and Greek and the clas-sics for two years, I also wanted to do science. As Iwas young, I could stay there in the sixth form for
Institute, 6701 Rockledge Drive, Bethesda, Maryland20892-7913, USA e-mail: [email protected].
This is an electronic reprint of the original articlepublished by the Institute of Mathematical Statistics in
Statistical Science , 2015, Vol. 30, No. 2, 258–267. Thisreprint differs from the original in pagination andtypographic detail. four years, mixing 2 years of classics (Latin, Greek,a little French and ancient history) with 2 years ofbiology, physics and chemistry. I never studied cal-culus, highly unusual in the US, but not so unusualat that time for a science student in England. WhenI got to the states, I estimated that what we did inEngland in the sixth form was equivalent to one ortwo years of undergraduate work in the states.
2. CAMBRIDGE UNIVERSITY
Gang:
How did you get to Cambridge University?
Robert:
In those days, the way you got to a uni-versity was either you were rich or royalty, or yousat for a scholarship examination. I applied to bothReading and Cambridge Universities and had totake a scholarship examination at each. The scholar-ship wasn’t much in terms of the money, but if youpassed the scholarship exam, the local governmentwould pay for your education.
Zhaohai:
So you got the scholarship?
Robert:
No, I actually failed. For Reading Univer-sity, the three subjects I chose for the exam wereGreek, French and chemistry. They thought thatcombination quite useless. At Cambridge, there weresix of us competing for one scholarship. At thattime, the School of Agriculture at Cambridge hada three-year Bachelor’s degree in agriculture. I hadalready decided to do something in agriculture. ButI had taken a special scholarship exam at Magda-lene College for people who would spend four years.The first two years would be Part I of the natu-ral science tripos. [A tripos is the course system atthe University of Cambridge.] Then at the end ofthe second year, I would have a choice, either a twoyear diploma in agriculture or continue with Part IIof the natural science tripos followed by a one yeardiploma in agricultural science. Although I didn’tget that scholarship, I did well in the exam, and theysaid they would accept me into Magdalene Collegefor the four-year program in two years’ time, withthe government giving me some support. Why twoyears? Because they “knew” I would have to servetwo years in the military. So I had those two years tospend. Since I was going to study agriculture, I wasable to get a deferment from the military to work ona farm, but I remained eligible for military serviceand could be called up later. So I worked on a farmfor a year and then spent my next year in Francewhere I perfected my French (my mother’s mothertongue).
CONVERSATION WITH ROBERT C. ELSTON Gang:
What happened after those two years?
Robert:
I returned to England and went to Cam-bridge. For the natural science tripos, I spent twoyears doing work for Part I. I had to have three sci-ence subjects. My original idea was to do botany,zoology and chemistry (organic and biochemistry).After one year, I really didn’t like botany and de-cided I wanted to do mathematics. So I changedfrom full subject botany to half subject botany andhalf subject mathematics. I had to teach myself cal-culus, which I did with a little book called
Calcu-lus Made Easy by Silvanus P. Thompson (1946).Clearly, I needed private tutoring for mathematics.Wally Smith (Walter Laws Smith), whom I knewfrom my extracurricular activity on the stage (wewere both members of the Pentacle club, which wasa magic club), became my mathematics supervisor.He told me I wasn’t very good at mathematics, butI stuck it out! I know mathematics lowered my examresult at Cambridge!
Zhaohai:
What did you do next? Did you get todo genetics at Cambridge?
Robert:
My first choice for a Part 2 tripos wasbiochemistry, which I really enjoyed, but betweenthe lectures and the labs, the hours were too long,so I went to my tutor for advice. I said I was think-ing about Part 2 genetics. I remembered his wordsso well, “You know this program in genetics here isnew. And this man [R. A.] Fisher is considered ec-centric by some, and it may not stand you in goodstead in later life for it to be known you workedwith him.” That is why I did not do a year withFisher! So I ended up doing the two-year diploma inagriculture.
Gang:
Who taught you statistics at Cambridge?
Robert:
I had lectures from four people. DennisLindley gave us three weeks on statistics as part ofthe half-subject mathematics. He taught me signifi-cance testing. He did not believe in it but he taughtit! In the same year, I did have lectures from Fisherbecause those who did zoology could do an optionalseries of lectures with Fisher on genetics. So I diddo genetics with Fisher. I’ll tell you a joke he told(translated into American). I did not know whathe was doing with adding, subtracting and dividingfor a 2 by 2 table. He came up with this number.He said, “Now, I am going to call this number chi-squared. Don’t be alarmed. I know you are all biolo-gists. It is no worse than calling a dog ‘Lassie’ !” Hesaid, “If this number is greater than 4, perhaps thereis something going on.” This was in a lecture hall which could hold 200 people. At the first lecture,there might have been 150 people; second lecture,50; third, about 15! He made us all sit in the frontrow. He could sense his audience. I don’t rememberexactly what he said, but it was something like: if Isay something is always transmitted from mother todaughter, then clearly it is never transmitted frommother to son. Then he said, “I do hope I’m notmaking a mistake in logic. Do stop me if I make amistake in logic, won’t you?” He could sense thatwe were all thinking about the truth of what he hadsaid!While I was spending a year in France, I had readFisher’s book for research workers (1950) and hisbook on experimental design (1951). So one day af-ter lecture I said to Fisher, I read your book. Whatis the difference between a standard deviation anda standard error? He looked me up and down andsaid: “Your height is a deviation from the mean. Itis not an error.”Then I also had lectures with Anscombe and R. C.Campbell, because they were teaching agriculturestudents. There we learned about experimental andsplit-plot design, and basic statistics. We also knewhow to calculate F-statistics using a hand calculator.And I had lectures from Wishart. Wishart used alittle book he had written with Sanders called
Prin-ciples and Practice of Field Experimentation (1955).He taught us how to lay out plots in the field andfor agricultural experimental designs. We were justagriculture people! He was in the School of Agricul-ture.
Zhaohai:
And you still wanted to be a farmer afteryou got your diploma in agriculture?
Robert:
What else was I going to do? I learnedhow to run a farm in England and knew quite a lotof animal physiology, plant physiology, soil science,how to work out the feeding and animal nutrition,but I had no capital. To be a farmer, you neededcapital for the land and the machinery. I couldn’tafford a farm.
3. COMING TO THE US FOR A PH.D.
Zhaohai:
How did you end up coming to US foryour Ph.D. in animal breeding?
Robert:
I got the B.A. in 1955. The way it workedwas that two years after you got your B.A., youcould pay 10 guineas and you got an M.A. So I hadan M.A. My mother wanted me to take an academicjob. I saw a notice: Fellowships to America. All right,
G. ZHENG, Z. LI AND N. L. GELLER
I thought, I’ll just go to America for just one year.These were King George VI memorial fellowshipsfrom the English Speaking Union of the US and theywere giving about 25 scholarships a year. You couldhave up to three choices of where you wanted to go,but you had to sign that you would go wherever theysent you. My choices were UC Davis where MichaelPease was doing chicken breeding or Ames, Iowa,which was known for dairy cow breeding. I left thethird choice blank. They sent me to Cornell, wherethere was a department of animal husbandry. ThereI was sent to Chuck (Charles Roy) Henderson, whosaid all his students minored in biometry, and sug-gested that I go to see Professor Federer. So I wentto see Walter Federer, who asked me why I did notstay for a Ph.D. I said I had only money for a year.He told me that they would find me money.I had to return to England at the end of the yearbecause I was called for military service; and if Ipassed my 26th birthday outside of England, I couldhave been called up to age 36. Again I avoided mil-itary service, this time by working on a pig farm. Iwas able to leave England before my 26th birthdaybecause the farm owner was willing to say I was stillthere. I actually spent my 26th birthday on the highseas en route back to Cornell. I did this so I couldget to Cornell when the semester began.So I returned to study animal breeding for a Ph.D.with Chuck Henderson with minors in biometryand mathematics. My thesis was in mixed modelnonorthogonal ANOVA. We had one of the firstcomputers, an IBM 650. I spent three months withpunch cards to invert a 79 ×
79 matrix!
Zhaohai:
Who were your contemporaries in grad-uate school at that time?
Robert:
I was exactly contemporaneous with astudent of Chuck Henderson’s, Shayle Searle. He hada degree in mathematics and a diploma in statis-tics. Chuck Henderson had just spent a year at theNew Zealand Dairy Board with Shayle Searle andrecruited him to be a graduate student at Cornell.I learned a lot of from both of them. Chuck Hen-derson told me he was an animal scientist, not astatistician. The reason he was doing BLUP (BestLinear Unbiased Prediction) was because statisti-cians wouldn’t do it for him. He never consideredhimself as a statistician at all.
4. FROM CORNELL TO UNIVERSITY OFNORTH CAROLINA AT CHAPEL HILL (UNC)
Gang:
How did you choose a post-doctoral fellow-ship in statistics after finishing up your Ph.D. inanimal breeding?
Robert:
I was going to finish my Ph.D. in thesummer of 1959. I didn’t know what to do next.Walter Federer advised me to do a post-doctoralfellowship in statistics. He said I should apply toPrinceton, where there were fellowships in statisticsfor biologists. They paid $5000 a year, tax free. SoI applied. Before I heard from Princeton, I drove toMiami for an international student conference. Onmy way back, I stopped at Chapel Hill to see myold friend Wally Smith, who had moved to the De-partment of Statistics there. He told me they couldoffer me $4,800 a year as a tax-free fellowship inthe Department of Biostatistics, as Bernard Green-berg, the chairman, had some money. So I went toUNC at Chapel Hill. I wrote to Princeton that Iwas no longer interested, and received the nicest let-ter back from Sam Wilks. From that experience, Ilearned that when you are trying to recruit studentsor post-docs, write a nice letter.
Zhaohai:
How long was your post-doc at ChapelHill?
Robert:
Just one year. Bernie Greenberg insistedthat teaching was part of the training for all stu-dents, pre- and post-docs, so I became a teaching as-sistant for Statistics 101 for public health. My affilia-tion was with biostatistics, but I had my office in thestatistics department. In addition, I took 5 theoreti-cal statistics courses in each of the two semesters, al-though I did not do all the homework. I took coursesin multivariate analysis from S. N. Roy and Nor-man Johnson, response surface designs from R. C.Bose, experimental design from Indra Chakravartiand David Duncan, nonparametric statistics fromWassily Hoeffding (U-statistics) and mathematicsfor statistics from Wally Smith. During that time,James Durbin, Maurice Kendall and E. J. Hannan(time series) were visitors. I also published workfrom my dissertation, my first paper in
Biometrics (1961).
Gang:
You spent most of 1959 to 1979 at UNCChapel Hill. Since you came as a post-doctoral fel-low, how did you manage to stay?
Robert:
In order for me to stay at Chapel Hill,Bernie Greenberg suggested a job as a Research As-sistant Professor of Pathology to work on a project
CONVERSATION WITH ROBERT C. ELSTON Fig. 1.
The department of biostatistics at UNC, Chapel Hill circa 1960. Front row (left to right): Thomas Donnelly, BernardPasternack, Robert Elston. Back row (left to right): Roy Kuebler, Jr., James Grizzle, David Newell (visiting), Bernard Green-berg, Bradley Wells. for the blood bank, which he thought must be re-lated to genetics, my interest. I worked on a projectto estimate the amount of blood that the blood bankshould keep on hand (Elston, 1962; Elston and Pick-rel, 1963). I simulated blood units being purchasedby the blood bank and being sent out for use (El-ston, 1962).Computers and statistics had not been used be-fore in blood banking. Analyzing six months of theblood bank’s records, I found I could fit a negativebinomial distribution to the number of units thatcame into the bank for seven of the eight majorblood types, but not for the O negative blood data.That was because O negative blood can be trans-fused into anyone and so O negative donors wereoften requested to donate blood, rather than do-nating simply at random. The blood bank directorwas impressed with that finding purely by statisticalanalysis. I ended up writing several other papers onthe blood bank project (Elston and Pickrel, 1965;Elston, 1966, 1968, 1970).Toward the end of that second year, Bernie Green-berg said he needed someone to teach bioassay thenext year, and did I know anything about it? I saidI knew a little, and yes, I could teach that, but intruth, I was one chapter ahead of the students mostof the time.
Gang:
But then there was this two year gap, 1962–64 when you went to Aberdeen. Why was that?
Robert:
My third year at Chapel Hill was the sixthyear I had been in the US. I had a J-1 visa and USlaw required that you had to return to your homecountry for at least two years. Exceptions to stayin the US were only by an act of Congress. DavidFinney contacted me for a permanent position as aSenior Biometric Fellow in Aberdeen Scotland and,since I couldn’t stay in the US, I accepted. My wifeand I didn’t like the idea of going to Aberdeen verymuch, but this was a permanent position. (I hadjust got married in Chapel Hill and my wife camefrom Gloucestershire, 100 miles west of London, sowe were both from southern England.) After beingin Aberdeen for about six months, I put down adeposit on a house, and the next day I got a let-ter from Bernie Greenberg asking me to come backto Biostatistics as an Associate Professor after therequired two years outside the US. He asked mewhat salary I would want and when I named thelargest salary I dared, he offered me 25% more. Iwas trapped! This time I came to the US with agreen card. We returned to Chapel Hill with a nine-month old daughter, and for her to get her visa, Ihad to sign on her behalf that she wasn’t coming
G. ZHENG, Z. LI AND N. L. GELLER into the US for the purpose of becoming a prosti-tute; she remarked recently that she kept her half ofthe bargain!
Gang:
Describe the Biostatistics Department onyour return.
Robert:
I was the 6th or 7th faculty member ofthe department. The department grew with the helpof federal grant support. In the mid-sixties, we weretremendously successful. In 1966, I managed to geta five-year Career Development Award and then afive-year renewal. There was an interdepartmentaltraining grant in genetics and biostatistics had itsown training grant, but there was no Ph.D. in bio-statistics. At that time, the Ph.D. students fundedby the departmental training grant took either aPh.D. in Experimental Statistics at Raleigh or aPh.D. in Statistics at Chapel Hill, with a minor inPublic Health. I wanted a Ph.D. program in bio-statistics with a minor in genetics to have studentsfunded by the interdepartmental training grant, soI wrote the Ph.D. proposal. Greenberg was told itwasn’t broad enough, so I rewrote it allowing forminors in genetics, demography and other fields aswell. The Ph.D. program in Biostatistics officiallybegan in 1968 with Rose Gaines-Das being the firstto get a Ph.D. in biostatistics, with a thesis in sta-tistical genetics.Initially, it was difficult for my students to getpositions. That’s why Joe Haseman went to the Na-tional Institute of Environmental Health Sciences;there were no academic positions in statistical ge-netics. Haseman could not get a job in statisticalgenetics despite the fact that Haseman and Elston(
Behavior Genetics , 1972, from Haseman’s disserta-tion) became the most cited paper ever published in
Behavior Genetics .By now, many of my Ph.D. students funded onthat training genetics grant are retired. I can’t imag-ine why!
Gang:
While you were at UNC, you did a lot oftraveling. How did you manage that?
Robert:
With the Career Development Award, myposition didn’t cost the university and Bernie Green-berg said I could do whatever I wanted because itdidn’t cost him anything. This allowed me to visitthe University of Hawaii to work with Newton Mor-ton and D. C. Rao for one year and, during sum-mers I had further trips to Hawaii and England (theGalton Laboratory in London and the Universityof Cambridge). When I visited the Galton labora-tory (1967), I met John Stewart, who was a grad-uate student at Cambridge. We ended up writing a paper together, which appeared in
Human Heredity (1971), a minor journal at that time. We computedthe likelihood of the model for the observed pheno-type data in a given pedigree. We could handle largepedigrees and relatively few markers. I didn’t thenknow I was using Bayes’ theorem recursively to com-pute the likelihood. Stewart’s contribution was toapply the result to linkage. In the discussion, Stew-art wrote that this paper answered a fundamentalquestion in human genetics, that is, is some phe-notype polygenic or is there a major gene? It wasKen Lange who named this “the Elston–Stewart al-gorithm.” It was overall a most productive time.
5. FROM UNC TO LOUISIANA STATEUNIVERSITY MEDICAL CENTER (LSUMC)
Gang:
Why did you leave UNC for LSU in 1979?
Robert:
I moved to LSU for two major reasons.I went to New Orleans for the ENAR meeting andthey wanted me to come there to be chair of theDepartment of Biometry in the LSU Medical Cen-ter. They offered me a hard money position. Thatwas the first reason: all positions at UNC were softmoney and by then I had four children, all withinsix years of age, to put through college. The secondreason was that at UNC I had gotten a grant whichallowed me to purchase my own computer and theuniversity would only let me house it in the com-puter center. At that time, nobody in the School ofPublic Health was permitted to have his own com-puter. So those were the primary reasons that I left.
Zhaohai:
Tell us about your years at LSU.
Robert:
Even though I had a hard money position,I kept writing grants. Because I had all of these fed-eral grants, I was able to start a Ph.D. program instatistical genetics and expand the faculty. I wrotefour proposals for Ph.D.s and masters’ degrees inBiometry and Genetics. Alec Wilson, Joan Bailey-Wilson and George Bonney became part of my fac-ulty.
Gang:
What kind of training did you give at LSU?
Robert:
I trained several post-docs there. I espe-cially like to train statisticians to do genetics. AtLSU, I had a training grant from NHLBI which wasinitially only to train post-docs. From 1992–1993,Dan Schaid of the Mayo Clinic was my post-doc. Iremember the year because Hurricane Andrew hitLouisiana that year. It was supposed to hit NewOrleans. We boarded up the windows and left forour eldest daughter’s wedding in Ann Arbor. We
CONVERSATION WITH ROBERT C. ELSTON thought we might not have a house when we gotback. But the hurricane missed New Orleans. DanSchaid completed his year with me and went backto the Mayo Clinic and was able to analyze the ge-netic data that they had been collecting. He is nowa leader in the field of genetic epidemiology. Zhaohai:
Why did you leave LSU?
Robert : My faculty was good and I wanted to raisetheir salaries. The administration said there wereno faculty raises and “no exceptions.” Of course,there were exceptions! That’s why I’m a lousy ad-ministrator: I refuse to lie! In one of the followingyears, the Chancellor wrote a letter to the depart-ment heads saying that again there would be no fac-ulty raises and noted that good people would leaveand “this should be taken as an opportunity.” I wasfed up with being a department head anyway andhad $1,000,000 in grant money. That and the cli-mate were the reasons I decided to leave!
6. FROM LSU TO CASE WESTERN RESERVEUNIVERSITY (CWRU)
Zhaohai : You have been at CWRU since 1995.Why did you choose CWRU?
Robert:
My wife hates the heat, so staying in thesouth was out of the question. She wanted to go toMaine or Vermont, so Cleveland was a compromise.I accepted a full professorship at CWRU withoutadministrative responsibilities so I could get somework done!
Gang:
Tell us about the department when you ar-rived.
Robert:
I was hired by an epidemiologist, AlfredRimm. He wanted me to have my own division, so wecalled it Genetic and Molecular Epidemiology. Thedepartment is really a mini-school of Public Health.Aside from my division, it had divisions of Epi-demiology, Biostatistics, Health Services Researchand Public Health. The names have changed overthe years, but with the exception that there are nolonger formal divisions, the structure is the same.When I moved, only two people from LSU camewith me, Xiuqing Guo, a graduate student, and He-mant Tiwari, a post-doc. I was also able to takemy training grant in biometric genetic analysis be-cause nobody remained at LSU who could do thework. Joan Bailey-Wilson and Alec Wilson movedto the National Human Genome Research Institute(NHGRI) of NIH because they had family nearby,in Baltimore. I was also able to take my computers.
Fig. 2.
Faculty of the Dept. of Biometry and Genetics, LSUMC, 1990. Front row (left to right): Bronya Keats, RobertElston, Joan Bailey-Wilson, Alexander Wilson, Mary Kay Pelias. Back row (left to right): Miguel (Mike) Guzman, VargheseGeorge, William Johnson, Yogesh Patel.
G. ZHENG, Z. LI AND N. L. GELLER
Al Rimm asked me to do genetics only, notbiostatistics and I did that for over ten years.One project was S.A.G.E. (Statistical Analysis forGenetic Epidemiology), which I had started in NewOrleans, funded by an NIH Resource Grant. I alsotook that with me. The Resource Grant required col-laborations, providing a service for which you hadto charge (the S.A.G.E. software), training and dis-semination (S.A.G.E. courses). Initially, there wasa charge for S.A.G.E. because the grant requiredthat. Beginning in 2005, we were able to distributeS.A.G.E. for free (see Elston and Gray-McGuire,2004 and https://code.google.com/p/opensage/ ).Version 6.2 was meant to be web based so peoplecould use other programs with it, but funding tocomplete this project never materialized.
Gang:
How come you became department chair atCWRU?
Robert:
In 2008, Al Rimm resigned as chair andthey asked me to be interim department chair and I agreed. They needed a real chair to apply for stimu-lus money, so in 2009, they took the “interim” away.They continued to advertise for a real chair and ittook several years—until now (2014)—to fill the po-sition. Call it a second childhood!
Zhaohai:
How did you arrange your time as chairon administration, research and mentoring graduatestudents and post-docs?
Robert:
When I first became chair, we needed toreorganize our Ph.D. program. Under the previouschair, the department was acting as though each di-vision had a separate Ph.D. program. The differentdivisions found it hard to agree on one Ph.D., butthe graduate school did not recognize multiple Ph.D.programs in one department. So I put a lot of ef-fort into establishing the one Ph.D. program in Epi-demiology and Biostatistics, with several differentconcentrations. When that was done, with severalfaculty committees to make sure that the programran smoothly, a lot of the administration was takencare of. And I have always considered my research
Fig. 3.
Professor Elston with students in the Division of Molecular and Human Genetics, Department of Epidemiology andBiostatistics, CRWU, 2004, with British Museum replica of the Rosetta Stone in the background. The caption above reads,“Genetic and molecular epidemiology is the Rosetta stone of medicine.” Front row (left to right): Kijoung Song, DeniseDaley, James Schick. Back row (left to right): Ronald Blanton, Murielle Bochud, Robert Elston, Danghong Song, CourtneyGray-McGuire.
CONVERSATION WITH ROBERT C. ELSTON Fig. 5.
Students with Professor Elston at a meeting of the International Genetic Epidemiology Society in Boston, August,2007. Left to right: Wei Guo, Qing (Jerry) Lu, Lu Zhang, Robert Elston, Xiaofeng Zhu, Xuefeng (Peter) Wang, XiangqingSun. and mentoring of students, both pre- and post-docs,to be all part of one and the same thing.These days I spend a lot of time writing grants.It’s getting harder to get them. Renewal of my re-search grants and the training grant is taking moretime than I wish. It’s hard to know how much de-tail the reviewers want. Sometimes the projects de-scribed in the grant proposal get published beforethe grant gets funded! And of course, if we don’t getfunding, we won’t be able to support Ph.D. studentsto do research.I also spend a lot of time helping others. For juniorfaculty, I tell them not to put my name on theirpaper as an author because when they come up fortenure, people may think the paper was my idea andnot theirs. If my name is on a paper, you can besure I really contributed something. I don’t noticeall the authors when I am reading a paper, but I findthat people notice if my name is on a paper. So Ihave to be sure that every sentence is accurate. Mypurpose is to be pedagogical as well as do research, and this makes me very fussy about proper wordingand clarity. I read the galley proofs personally and Ihave my secretary read them, too. I learned that atAberdeen from Finney. He had a sign up in the tearoom which said, “No paper leaves this departmentwithout the Professor’s permission.”I still work on family studies, although that hasbecome less fashionable than case-control GWAS.My recent work is still a mixture of theoretical andapplied and I still enjoy writing and publishing withstudents.
Gang:
Do you ever plan to retire?
Robert:
I don’t know when I shall retire—probablywhen I am no longer able to get grants to fund myresearch. I have four children and ten grandchildren.By the way, three of my children are university pro-fessors in mathematics/health sciences, and the onewho isn’t decided to be creative and studied acting.She puts on the high school play every year (and herolder son is majoring in mathematics at college). I G. ZHENG, Z. LI AND N. L. GELLER
Fig. 4.
Robert Elston at Case Western Reserve University,2007. look forward to spending more time with my familywhen I do retire.
7. SUMMING UP
Gang : You have directed 40 Ph.D. theses and had45 post-docs. By now they too have had trainees.What does your “research pedigree” look like?
Robert:
My research pedigree has more than 500progeny The International Genetic EpidemiologySociety had a special tribute for me on my 70thbirthday and someone drew it out. At that time,half of the field of genetic epidemiology was in mypedigree. There are at least four generations.
Gang:
How would you sum up your career?
Robert:
Like that of many other academics, mycareer path was an accident. When I talk to othersin academia, most of the time they had no idea whatfield they would end up in. In my case, I didn’t evenexpect to go to academia. But once I decided toapply statistics to genetics, I think I made a happychoice; and I’m glad I decided early on to make all ofmy students collaborators. I only hope they learnedas much from me as I from them.
Gang:
What is your advice to a young statisticalgeneticist starting out today?
Robert:
My advice is quite simple. First, makesure you keep learning as much statistics as you canand second, keep up to date with computing tech-nology. Statistical genetics may go out of fashion,but there will always be a need for statisticians whocan compute.
Zhaohai:
Do you have any closing comments?
Robert:
It is really nice to have this conversationappear in
Statistical Science . I actually never con-sidered myself to be a statistician. I was a geneticistamong the statisticians and a statistician among thegeneticists! REFERENCES
Elston, R. C. (1961). On additivity in the analysis of vari-ance.
Biometrics Elston, R. C. (1962). Simulation d’un Processus Stochas-tique Impliqu´e dans la Gestion d’une Banque de Sang.
Biometrie-Praximetrie Elston, R. C. (1966). Letter to the Editor: Computers andBlood Bank control.
J. Am. Med. Assoc.
Elston, R. C. (1968). Inventory levels for a hospital bloodbank under the assumption of 28-day shelf life.
Transfusion Elston, R. C. (1970). Blood bank inventories.
CRC CriticalRev. Clin. Lab. Sci. Elston, R. C. and
Gray-McGuire, C. (2004). A reviewof the “Statistical Analysis for Genetic Epidemiology”(S.A.G.E.) software program package.
Hum. Genomics Elston, R. C. and
Pickrel, J. C. (1963). A statistical ap-proach to ordering and usage policies for a hospital bloodbank.
Transfusion Elston, R. C. and
Pickrel, J. C. (1965). Guides to inven-tory levels for a hospital blood bank determined by elec-tronic computer simulation.
Transfusion Elston, R. C. and
Stewart, J. (1971). A general modelfor the genetic analysis of pedigree data.
Hum. Hered. Fisher, R. A. (1950).
Statistical Methods for Research Work-ers , 11th ed. Oliver and Boyd, Edinburgh.
Fisher, R. A. (1951).
The Design of Experiments , 6th ed.Oliver and Boyd, Edinburgh.
Haseman, J. K. and
Elston, R. C. (1972). The investigationof linkage between a quantitative trait and a marker locus.
Behav. Genet. Thompson, S. P. (1946).
Calculus Made Easy , 3rd ed.Macmillan, London.
Wishart, J. and
Sanders, H. G. (1955).