how can outliers be removed from data before making a graph?

This is actually a very controversial question. If you manually choose which experiments "just don't look right," you are at risk of censoring your data into the shape that confirms your hypothesis. You don't want to fool yourself with your own data.

Instead, I would look to metrics like Tukey's Outlier rule: "outliers are values more than 1.5 times the interquartile range from the quartiles — either below Q1 − 1.5IQR, or above Q3 + 1.5IQR."