> For the complete documentation index, see [llms.txt](https://lwm.mit.edu/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://lwm.mit.edu/course-materials/computation-questions/module-2-screen-to-signal.md).

# Module 2: Screen to Signal

Unlocking Functional Targets with AI and Machine Learning

**Date:** Friday, October 9, 2026 (12-1:30 PM)

**Presenters:** Sam Maffa (Li Lab, Whitehead) and Alexander Greer (Vander Heiden Lab, KI)

## Overview

{% hint style="info" %}
**Central Question: "**How can machine learning analyze screening data to identify biological dependencies?"

**Learning Objectives:**

* Understand the structural challenges associated with examining high dimensional data
* Learn how Random Forest classifiers can extract trends from data
* Interpret feature importance plots and predictability scores
* Connect screening results to biological knowledge to identify mechanism
  {% endhint %}

<table data-view="cards"><thead><tr><th></th><th data-hidden data-card-cover data-type="image">Cover image</th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><h4>Presentation Slides</h4></td><td><a href="https://2078575508-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FpK8DMWsoyo5vzRk1L6UL%2Fuploads%2FjGBjJi5gqAG56SEZcFyR%2Fm2_cover.png?alt=media&amp;token=f8f640f8-167c-49af-9b9a-d919316f65eb">m2_cover.png</a></td><td><a href="https://docs.google.com/presentation/d/1wijUKrWCvMrfP99tK1EIz-nDDs7VG7hIicQVJ6YizTw/edit?usp=sharing">https://docs.google.com/presentation/d/1wijUKrWCvMrfP99tK1EIz-nDDs7VG7hIicQVJ6YizTw/edit?usp=sharing</a></td></tr><tr><td><h4>Code Notebooks</h4></td><td><a href="https://2078575508-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FpK8DMWsoyo5vzRk1L6UL%2Fuploads%2FFTkJDrRVoYQm9FfUne9R%2F1765109208272.png?alt=media&amp;token=27836f71-a47a-49b8-acf1-a5c2e46a8129">1765109208272.png</a></td><td><a href="https://drive.google.com/drive/folders/1dy_RK7Uu7NR61IuxdCBuXMVEdn5WYl31?usp=sharing">https://drive.google.com/drive/folders/1dy_RK7Uu7NR61IuxdCBuXMVEdn5WYl31?usp=sharing</a></td></tr><tr><td><h4>Reference Guide</h4></td><td><a href="https://2078575508-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FpK8DMWsoyo5vzRk1L6UL%2Fuploads%2FkqDoJFZYBOWo11m38vMr%2Fchoose-markdown.jpg?alt=media&amp;token=ac805070-e315-4d72-8f54-4bcaa5376197">choose-markdown.jpg</a></td><td></td></tr></tbody></table>

## Motivating Problem

High throughput experimental methods generate vast quantities of relevant biological data. But human analysis cannot alone analyze trends and behaviors across so many dimensions. Computational tools permit us to systematically explore high dimensional data to identify information, patterns, and dependnecies potentially relevant to our biological question.

## Exercise

This session's exercise uses computational tools to load and analyze high dimensional screening data relevant to a biological question. The code workbooks employ random forest models to segment the data to reveal cell line-specific dependencies.

The notebooks from this session use data and methods adapted from:&#x20;

* [*WRN helicase is a synthetic lethal target in microsatellite unstable cancers*](https://doi.org/10.1038/s41586-019-1102-x)
* [*SKI complex loss renders 9p21.3-deleted or MSI-H cancers dependent on PELO*](https://doi.org/10.1038/s41586-024-08509-3)

## Discussion <a href="#discussion" id="discussion"></a>

Ask questions and share your own applications of these techniques in the Collaboration Park (see the **Computation Questions** channel):

{% content-ref url="/pages/nXkZZhFpJGTQPmWCg9j4" %}
[Community](/community.md)
{% endcontent-ref %}