liveBook · Manning

Interquartile range

As explained in Build AI Drug Discovery Pipelines

The interquartile range is the difference between the third and first quartiles of a distribution.

The interquartile range (IQR) measures the spread of the middle half of solubility values: IQR = Q3 − Q1, where Q1 and Q3 are the 25th and 75th percentiles. section 4.1.2

Use the IQR rule to flag potential solubility outliers:

  • Calculate Q1, Q3, and IQR from the solubility values. section 4.1.2
  • Flag values below Q1 − 1.5 × IQR or above Q3 + 1.5 × IQR. section 4.1.2
  • Inspect flagged molecules manually before removing them, because extreme values may represent legitimate members of the target population. section 4.1.2

Solubility data cleaning

Extremely insoluble molecules, metal-containing entries, and hydrogen fluoride may fall outside the intended domain of organic small-molecule drug candidates; chemical knowledge complements the IQR rule. section 4.1.2

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Defines quartiles, computes IQR, specifies 1.5-IQR thresholds, and defers final decisions to manual and chemical inspection.
Examples
Uses extreme solubility values, metal-containing molecules, hydrogen fluoride, and rendered molecule comparisons.
What it assumes
Assumes solubility values and percentiles, while defining Q1, Q3, and their role in the calculation.
What you can do after
Diagnose potential solubility outliers and combine statistical flags with chemical judgment before filtering training data.
How it connects
Connects quartiles and percentiles with outlier detection, solubility prediction, chemical domain knowledge, and data preprocessing.

Also explained in

Compare how 8 books explain interquartile range →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build AI Drug Discovery Pipelines. The section links point to the parts of the book it draws on.