SHRINIWAS K.

KATTI Name and address: Shriniwas K Katti 221 High Point LN Raymore MO 64083-9096 Contact Information Current Home Telephone: 816 348 9046 Cell Phone 816 506 8889 Email (Permanent): sk13b6c7c@westpost.net Citizenship: USA Naturalized in 1970. COMPUTER HARDWARE AND SOFTWARE EXPERTISE Hardware: APL Computers: PC, VAX, UNIX, Main frame on IBM. Data Management Experience: I was located next to the data management group which was given the job of putti ng the data that came from many forms onto the computer. They also would find er rors in the data sets and they would send queries to the nursing staff and the M edial Doctors to get the data set straightened out. I would take it from there a nd check the data for errors that they Data Mangers would miss. This took about half of my time. After that came the job of SAS programming to get the tables, listings graphs an d statistical analyses. Programming Languages: Visual Basic, Pascal, SAS Interaction with database: SQL Software Packages: SAS, MINITAB, IMSL, Scientific Word, WordPerfect, Excel EDUCATION Ph.D. Statistics, Iowa State University, Ames, Iowa M. A. Statistics, Iowa State University, Ames Iowa B.Sc. Mathematics, University of Delhi, Delhi, India PROFESSIONAL EXPERIENCE Post 2008 Experience As a retiree with many contacts at all levels, I continued to work with the MDs and my associates. When we make many tests of hypotheses, there is a multiple test with the exact pre-set Type I error under the condition that the estimates being used are indep endent and the variances have enough large degrees of freedom so that their use in the multiple test does not increase the Type I error. So, one does not have t o use the conservative test used in SAS. There is an exact test for multiple com parisons. An MD, Dr. Nenad Markovic of NIH, and I are working on a screening test for you ng girls which has high power and specificity. We convinced the FDA that we do n ot need sample sizes of the order of 300. Specificity is the hard part. But we c an detect excessive growth with our method quite easily. Associating our test sp ecifically to cancer of the cervix must use other considerations. An MD from NSF and I are trying to set up a method to market the protocols and devices that have been developed in their laboratories in India. My family in In

dia is quite ready. July 2008 _ August 22, 2008: SDC Clinical *Statistics and Data Corporation, Developing TLGs for many rather simple trials. Jan. 3 2008 April 30, 2008: Statistician, Premier Research Group, Atlanta branc h but working from home in Raymore, MO. The appointment is for six months with scope for permenancy. Reading up on sedation of youngsters (<16 years of age). Finished writing an extensive Training Manual to introduce the local staff to b iostatistics and lead them in developing protocols, SAP and final reports. Obtain sample sizes for two different studies. Write Statistical Analysis Plans for two projects. Sept. 24, 2007 October 18, 2007 Sr, Biostatistician, Sanofi Aventis Company, 3 00 Somerset Corporate BLVD, Bridge Water, NJ Read up on Diabetes. Analyzed data collected by the Managed Care Group from Ins urance Companies with focus on the Sanofi Aventis drugs. It was a data mining jo b. March 29, 2007 June 15, 2007 - SR Biostatistician, Bristol Myers and Squibb, N Billerica, MA, Reviewed two protocols, set up to test two drugs that contained their main pate nted drug, Definity. As written, they were cut and paste. By writing them from s cratch, one can make them more explicit and permit easy collection of data and a nalysis. Reviewed TLGs for two studies that had been stopped very early. The data was ve ry scrappy. So, while catching many errors, not much more could be done than to develop elementary Tables and Adverse Events; and put out listings. It was all d one in SAS. Attended a Dinner Meeting called by an MD to discuss the over all role of stati stics in Drug Trials. May 2006 - February 2007 (private contractor) Sr. Biostatistician, from home Work with Doctor Nenad Marcovic of Oncology who is in the process of developing an NDA for testing ovarian cancer. Seeing scattered raw data, it is quite an ef ficient method. Work with Dr. John Rothman to write a paper on a new algorithm aimed at physici ans to permit them to select the drug mot suited for that patient. We note that when Drug A is proven better than Drug B, we show that the average effect of Dru g A is higher than the average effect of Drug B. A given patient could be reacti ng better to Drug B than to Drug A. We have a recommendation on how a physician can do mini-trials on his patients and keep doing it to choose the best drug at any given time. November 2004 - April 2006, PAREXEL (CRO) All of the work was done in SAS Wrote SAP for a Phase I study. Helped determine the Maximum Tolerable Dose (Dose Limiting Toxicity) and the do se to be use in the next Phase II study. I fitted a logistic distribution to est imate the dose that has a probability of 1/3 of generating Dose Limiting Toxicit y. All intermediate calculations were done in SAS SAS is set up for such work. Act as a resource for other biostatisticians: I showed them what the likelihood

for the mixed model is and where the various parameters are located. Mentored s ome workers; gave general lectures. An interesting project was to find a formula that tells you when 30% of the can cer patients will be deceased based on current data to estimate the date when th e interim analysis is due. This was done in SAS and this program was made a par t of their data base so that when they add patients to the database and run it, they would automatically get an updated value for the Interim Analysis. Review and validate SAS programs and statistical reports before they go out of our Department. I would program in SAS to validate their work. October 2003 September 2004 SR Biostatistician (contract position), Eli Lilly P harmaceuticals, Greenfield, IN, Statistician, Animal Sciences Division FDA/DVM submissions and ICH Guidelines: o Project statistician to finish a submission to CVM. CVM approved our submissio n without any revision. The work had been initiated in SAS; so I continued with it in SAS. o Established that the Lilly Chlorine Pills have high stability 132 months. Pre vious statisticians were getting insignificance for the period of stability beca use the method used to estimate the amount of chlorine in the pill had high vari ability. There were cases where estimated amount was larger than the amount put into the pill. So, the model needed adjustment. This figure was accepted by the FDA to which this was submitted. All of the work was in SAS since FDA is equippe d with SAS to validate our work quickly. April 2003Aug. 2003, Quintiles (CRO), Kansas City, MO, Senior Statistical Scient ist Provided statistical expertise to the other groups mentoring and giving brief lectures on some topics that they had missed. An interesting aspect was to show them how to write a MACRO in a systematic way. Thus, write a program first and c heck it out. If you are going to use it again, convert it into a MACRO this is better than cutting and pasting as they were doing. Promoted better statistics through seminars, lecturing etc. Senior Statistical review of reports, protocols and Statistical Analysis Plans. While SAS was the central software, I used other matrix based software to compu te difficult formulae. Worked with programmers and junior statisticians in Phase I IV studies. I wou ld show them where the error, why it is wrong and how the correction is discover ed. All the programmers whom I mentored are still in the programming business. Areas of interest: Asthma, apnea, use of insulin in growth retardant population . May 2002Aug. 2002, Hoffman La Roche, Nutley, NJ, Consulting Biostatistician II Acted as a resource person and Lecturer of Research Statistics for the local st atisticians Reviewed the literature on multiple comparisons and showed that there is an exa ct test. Dec. 2001May 2002, Tyco/Mallinckrodt, St. Louis, MO, Consulting Bio-statistician Finished a Phase IV submission on imaging for cancer detection. My Due Diligenc e part consisted of drawing high density graphs that depicted the data in the va rious areas. Problems would become apparent which otherwise would be masked by the averaging that occurs in statistics. SAS/GRAPH is well suited for this work.

On rechecking, I found some unusually high numbers to be data transmission erro rs. Finished Meta Analysis of 28 papers on the use of two different agents for high lighting images. I had a well-defined recommendation. The imaging solution with a low half-life had immense superiority over those with long half-life. All work was done in SAS. March 2000Dec. 2001, Progenics Pharmaceuticals Inc., Tarrytown, NY, Consulting B io-statistician This was a small company headed by four medical doctors. They would make up dru gs based on their knowledge of the theses and dissertations in the major univers ities. Experiments in sets of four animals and three patients would be made at t he graduate school level. If any drugs showed effectiveness, it would be set up for the phase II trial with FDA, NDA and all. Three of the drugs that I selected are now going through Phase II. Cancer and interferon: I noticed that the vari ability was immensely higher than their expectations. My suggestion was that at every stage, three independent assays be made of the blood sample and any tissue samples. My main conclusion was: that Interferon should be phased out over a t hree day period. After that you must phase in immunity building drugs to high le vels. Stroke: Phase I study was made on baboons with the surgery being done at the Co lumbia University. It served the purpose of clarifying the issue of the level of drug to be used. It pointed us to the many errors that are possible when there is arterial occlusion. This also served to pin down an effective dose. The type of work we did, with four to ten patients, did not permit the use of S AS. I would derive formulae and compute them using Excel. HIV study: We had data on people in sets of five and ten with fairly strong HI V infection. They had many observations before dosing; then they had taken blood samples one hour post dosing, two hours post dosing, four hours post-dosing, ei ght hours post dosing, 16 hours post-dosing, 48 hours post-dosing and 96 hours p ost-dosing. The variability is so high that when you draw HIV load against time for each patient, the graph is berserk up and down that some of the better docto rs were concluding that we have not yet obtained a drug which has any effect of reducing HIV load. I developed a new method. I took al pre-dosing observations and computed the mean and standardized it by multiplying it by the square root o f the number used in computing the mean so that the variability of the standardi zed mean stays constant. I took the data at the four time periods and computed a similar standardized mean. Same with the next four observations. Then put toget her the data on all patients and all doses. The dose response-response relation was obvious. Based on this graph, I estimated the dose needed to reduce the HIV load by 90%. Between the biochemist and the Medical Doctor, they made additional trials on a few more patients and the data ell right on my dose-response curve. After we convinced ourselves of the existence of a dug to cure HIV, our Biochem ist and MD went to work with other personnel in better resourced industries to d evelop better drugs that are easier to develop; made trials and established a fa irly inexpensive method to reduce HIV by 50%. Soon, international trials were ma de and HIV was no more an unknown disease. It was just like any virology disease with high variability in detection. Read more about it in an article entitled, What I bring to the Table. October 1999Mar 2000, Janssen, Titusville, NJ, Consulting Bio-Statistician Wrote the statistical part of the protocols and gave mock-tables. I gave tests for equivalence and organized sequential tests. It was set up in S AS format. June 1999-Oct. 1999, Janssen, Titusville, NJ, Consulting Bio-Statistician Wrote the statistical part of the protocols and gave mock-tables.

I gave tests for equivalence and organized sequential tests. June 1999-Oct. 1999, ICON Inc. (CRO), Philadelphia, Consulting Bio-Statistician Created Tables and Graphs for the clients. List of clients included Novartis an d other major drug companies. All work was done in SAS. The original companies a nd FDA are all set up with SAS. So, gong along with SAS lead to speedy approval. Validated programs written by others. This is 100% SAS work. On time submissions and accuracy were the major issues. We always got the thing s done faster than the demand. So, we were given additional work with additional funding. I wrote flow chats and mini-SAS-programs for each piece. When a new job would c ome, we would see which piece apply. So, we did not hve to start from scratch of each ne set of TLGs. April 1999-June 1999, New Drugs Services, Kennett Square, PA, Statistician Review and revise Protocols, SAPs, Tables, Listings and Graphs. This was all ba sic SAS-work. Sept. 1998-March 1999, Eli Lilly Pharmaceuticals, Indianapolis, IN, Statistician Review and revise Protocols, SAPs, Tables, Listings and Graphs. This was all ba sic SAS-work. Dec. 1997-March 1998, Abbot Pharmaceuticals, Dallas, TX, Statistician Review and revise Protocols, SAPs, Tables, Listings and Graphs. This was all ba sic SAS-work. Sept. 1997-Dec. 1997, Florida International University, Miami, FL, Professor of Statistics Taught two course, gave seminars May 1996Dec. 1996, Ryder Trucking Company, Miami, FL, Statistician Use the APL-programming. They were trying out APL in comparison with SAS. The f inal decision was to use SAS. SAS is better set up for missing values. SPECIAL ACCOMPLISHMENTS: 85+ publications in statistical journals on methodologies on topics including 1. 2. 3. 4. 5. 6. 7. 8. 9. Design of Experiments Linear Models Non-Linear Models Multivariate Analysis Bayesian Methods Tests for Normality Independence/divisibility Medical/Biological Applications Computer Work

Fellow, American Statistical Association. Director of Graduate Studies in Statistics. Guided dissertations. Former Associate Editor of Biometrics. Founding Father of the Mid-Missouri Chapter of the ASA. Founding father of the Katti Family Endowment at UMC, value $100K.

UNIVERSITY Professor Professor Assistant 960-1970.

EXPERIENCE: of Florida International Univ., 9/97-12/97 of University of Missouri-Columbia, 1970-1995 and Associate Professor, Florida State University, Tallahassee, FL, 1

1. It was I who started the Graduate program at Florida State University, starti ng with getting it approved, hiring faculty and students and developing program and administering it. 2. I was the Director of Graduate Studies for the entire period of ten years tha t I was there. 3. At the University of Missouri, they already had the Graduate Program approved but they had no faculty to develop it and get the students with enough knowledg e to be given MA and Ph.D. in Statistics. I was hired and assigned the position of Full Professor and Director of Graduate Studies. I developed a program and di d every thing that a Director of Graduate Studies does to start giving students their MA and Ph.D./ in statistics. None of our students had ever any difficulty in getting employment. I was the Director of Graduate Studies for 20 of the 25 y ears that I was there. FOREIGN LANGUAGES: Many languages of India, including Hindi, Marathi and more; r eading language of Sanskrit LIST OF PUBLICATIONS is available. Please ask.