Figure 1

Test of equivalence for five groups.

## 
##  Kruskal-Wallis rank sum test
## 
## data:  siglec by cirrhosis_status
## Kruskal-Wallis chi-squared = 15.327, df = 4, p-value = 0.004068

If we group Healthy Controls and Sepsis together.

## 
##  Wilcoxon rank sum test with continuity correction
## 
## data:  siglec by group
## W = 1508.5, p-value = 0.0001088
## alternative hypothesis: true location shift is not equal to 0

Figure 2

Sick patient only.

## Call:
## coxph(formula = Surv(fucensor_days, death_90_admission) ~ siglec_bin, 
##     data = subset(dat3, cirrhosis_status != "Healthy Controls"))
## 
## 
##             coef exp(coef) se(coef)    z      p
## siglec_bin 0.741     2.098    0.281 2.64 0.0083
## 
## Likelihood ratio test=7.29  on 1 df, p=0.00693
## n= 128, number of events= 55

All patient. I added this in case reviewer request it.

## Call:
## coxph(formula = Surv(fucensor_days, death_90_admission) ~ siglec_bin, 
##     data = dat3)
## 
## 
##             coef exp(coef) se(coef)    z     p
## siglec_bin 0.808     2.242    0.281 2.88 0.004
## 
## Likelihood ratio test=8.67  on 1 df, p=0.00323
## n= 134, number of events= 55

Figure 3

Figure 4

Corrlation between siglec (original scale) and AST, ALT, APRI (after log-transformation).

##             siglec   log.ast   log.alt  log.apri
## siglec   1.0000000 0.3438546 0.2895265 0.5079886
## log.ast  0.3438546 1.0000000 0.8524903 0.8282806
## log.alt  0.2895265 0.8524903 1.0000000 0.7003898
## log.apri 0.5079886 0.8282806 0.7003898 1.0000000