Software for supporting large scale data processing for High Throughput Screening

Abstract

High Throughput Screening for is a valuable data generation technique for data driven knowledge discovery. Because the rate of data generation is so great, it is a challenge to cope with the demands of post experiment data analysis. This thesis presents three software solutions that I implemented in an attempt to alleviate this problem. The first is K-Screen, a Laboratory Information Management System designed to handle and visualize large High Throughput Screening datasets. K-Screen is being successfully used by the University of Kansas High Throughput Screening Laboratory to better organize and visualize their data. The next two algorithms are designed to accelerate the search times for chemical similarity searches using 1-dimensional fingerprints. The first algorithm balances information content in bit strings to attempt to find more optimal ordering and segmentation patterns for chemical fingerprints. The second algorithm eliminates redundant pruning calculations for large batch chemical similarity searches and shows a 250% improvement for the fastest current fingerprint search algorithm for large batch queries.