Internship: Protocol development for machine learning for

Internship: Protocol development for machine learning for phenotypic cellular analysis
“A picture is worth 1000 words” expresses the concept that there is a lot of information within images.
Using microscopy images of human cells in culture is a powerful tool for studying cell biogenesis and
function. Image features often change when cells are treated with compounds or siRNAs. Variation
among image features can be quantitated to better understand the effect of the treatment, but the
changes in the features are not always obvious to the eye. Machine learning is an analysis approach to
identify and quantify phenotypic differences among experimental cellular populations which can be
applied to large imaging data sets to discover distinct patterns among experimental groups.
Yale Center for Molecular Discovery is looking for a summer intern to develop our abilities in image
quantification and machine learning for existing data sets and software. In this internship,
comprehensive analysis of the available machine learning freeware (including CellProfilerAnalyst) will be
performed. You will learn and use CellProfilerAnalyst to analyze and quantify image features in image
data. Building on that, you will also develop machine learning techniques to make a list of morphology
differences between the experimental groups. Finally, time permitting, this list will be used for Principal
Component Analysis (PCA), a method which allows identifying the features that show the most
variation. All developed protocols and results will be documented for future use by YCMD and Yale
This is a great opportunity for a student focused on computational sciences. Some statistical
background is helpful but not required for this internship.