2. In the survival package, there is a data set called mgus which gives the result of a study of 241 subjects with monoclonal gammapothy of undetermined significance(MGUS). These subjects are at increased risk of developing certain blood diseases. We are interested in using the given covariates: age age of the patient when they were diagnosed, sex male or female, alb albumin level at MGUS diagnosis, creat creatinine level at MGUS diagnosis, hgb hemoglobin concentration at MGUS diagnosis, mspike size of the monoclonal protien spike at MGUS diagnosis. to model the survival time for these patients. We will use futime as our time-to-failure measurement and the variable death indicates whether the patient died or was censored. Using a Cox proportional hazards model and likelihood ratio tests, (a) Test the hypothesis that there is a significant difference between the two sexes. Report a P-value and a conclusion. (b) Re-run the test of the difference between the sexes, but use the covariates age, alb, creat, hgb, and mspike to control for differences in the groups. Report a P-value and a conclusion. (c) How might we explain the difference in the answers to part (a) and (b)? (d) Propose a set of covariates that you think best fits the futime data in a not-too-complicated way. Justify your choices with regression results