Slim-Filter: An interactive windows-based application for Illumina Genome Analyzer data assessment and manipulation
G. Golovko1,2, K. Khanipov1, M. Rojas1,2, A. Martinez-Alcántara1, J. J. Howard1, E. Ballesteros1, S. Gupta1, W. Widger1,3, and Y. Fofanov1,2,3
1Center for BioMedical and Environmental Genomics, University of Houston, Houston, TX, USA. Department of Computer Science2 and the Department of Biology and Biochemistry3, University of Houston, Houston, TX, USA 77204
The emergence of Next Generation Sequencing technologies has made it possible for individual investigators to generate gigabases of sequencing data per week. Effective analysis and manipulation of these data is limited due to large file sizes, so even simple tasks such as data filtration and quality assessment have to be performed in several steps. This requires (potentially problematic) interaction between the investigator and a bioinformatics/computational service provider. Furthermore, such services are often performed using specialized computational facilities.
We present a windows-based application, Slim-Filter designed to interactively examine the statistical properties of sequencing reads produced by Illumina Genome Analyzer and to perform a broad spectrum of data manipulation tasks including: filtration of low quality and low complexity reads; filtration of reads containing undesired subsequences (such as parts of adapters and PCR primers used during the sample and sequencing libraries preparation steps); excluding duplicated reads (while keeping each read's copy number information in a specialized data format); and sorting reads by copy numbers allowing for easy access and manual editing of the resulting files. Slim-Filter is organized as a sequence of windows summarizing the statistical properties of the reads. Each data manipulation step has roll-back abilities, allowing for return to previous steps of the data analysis process.Slim-Filter is written in C++ and is compatible with fasta, fastq, and specialized AS file formats.
Slim-Filter Performance was estimated using following computer configurations:
| Windows | Linux | |
| OS | Windows 2008 Server SP1 | CentOS 5.6 |
| CPU | Dual Quad Core Intel™ Xeon W5590 3.33GHz,8M L3 | AMD Magny Cours 6128 8-Core Processor, 2.0 GHz, 12MB Cache |
| RAM | 128GB, DDR3 RDIMM, 1066MHz, ECC | 512 GB DDR3 1333Mhz ECC |
| HDD | 2TB SATA 3.0Gb/s, 7200 RPM | 2TB SATA 3.0Gb/s, 7200 RPM |
| Number of reads 36 bases long | RAM required to perform computations for Linux (Mb) | RAM required to perform computations Windows (Mb) | Time to apply all possible filter settings in Windows (seconds) | Time to apply all possible filter settings in Linux (seconds) |
| 10,000 | 10 | 32-40 | <1 | <1 |
| 100,000 | 25 | 85-100 | 6.5 | 4.5 |
| 1,000,000 | 300 | 600-800 | 66 | 40 |
| 10,000,000 | 2,000 | 3,500 | 590 | 442 |
| 50,000,000 | 13,000 | 45,000 | 3,000 | 2,156 |
No comments:
Post a Comment