diff --git a/README.md b/README.md
index 98cf6f4..e7c96af 100644
--- a/README.md
+++ b/README.md
@@ -125,8 +125,8 @@ weiss2021_vanc %>%
 
 ![](tools/readme/weiss_vancomycin_index-1.png)<!-- -->
 
-For healthy children, the medain value of the index is about 0.2,
-whereas it is roughly 0.85 across the samples from children with sepsis.
+For healthy children, the medain value of the index is about -0.2,
+whereas it is roughly -0.85 across the samples from children with sepsis.
 Let’s look further into how the index was calculated, and check out
 which bacteria were labeled as susceptible or resistant to vancomycin.
 
@@ -246,7 +246,7 @@ Most taxa in the sample are annotated as susceptible to vancomycin,
 including the most abundant taxon, *Ruminococcaceae*. One taxon, RF39,
 is not annotated. Only a few taxa are annotated as resistant to
 vancomycin, thus it’s not surprising that the vancomycin index for the
-sample is negative.
+sample is positive.
 
 ``` r
 healthy6_data %>%
@@ -257,13 +257,13 @@ healthy6_data %>%
 #> 1 0.207
 ```
 
-How would we expect the proportions to change if the index increased to
-a positive value, say 0.5? We can use `predict_abundance()` to run the
+How would we expect the proportions to change if the index decreased to
+a negative value, say -0.5? We can use `predict_abundance()` to run the
 calculation.
 
 ``` r
 healthy6_data %>%
-  mutate(predicted = predict_abundance(0.5, proportion, susceptibility)) %>%
+  mutate(predicted = predict_abundance(-0.5, proportion, susceptibility)) %>%
   rename(observed = proportion) %>%
   pivot_longer(
     c(observed, predicted), names_to = "method", values_to = "abundance") %>%
@@ -283,11 +283,11 @@ the taxon that’s not annotated, RF39 (near the middle), the abundance
 has not changed at all.
 
 To finish, let’s re-calculate the vancomycin index for our predicted
-abundances, so we can verify that it has the expected value of 0.5.
+abundances, so we can verify that it has the expected value of -0.5.
 
 ``` r
 healthy6_data %>%
-  mutate(predicted = predict_abundance(0.5, proportion, susceptibility)) %>%
+  mutate(predicted = predict_abundance(-0.5, proportion, susceptibility)) %>%
   summarise(vanc = mirix_vancomycin(predicted, lineage))
 #> # A tibble: 1 × 1
 #>    vanc