Results: 4

1.
Figure 1

Figure 1. From: ReadDB Provides Efficient Storage for Mapped Short Reads.

ReadDB architecture. ReadDB uses a client-server architecture to give clients access to a large set of mapped read positions. The server stores each set of aligned reads in one directory and uses several files per chromosome. The index files allow quick access to the read data (which is sorted by position) and are small enough that the server can cache many index files in memory at once.

P Alexander Rolfe, et al. BMC Bioinformatics. 2011;12:278-278.
2.
Figure 3

Figure 3. From: ReadDB Provides Efficient Storage for Mapped Short Reads.

Performance comparison with no network latency. Comparison of runtimes with the client software and server on the same machine. To separate the effects of network throughput and latency from those of the disk and CPU, we ran the test suite from the server machine. The ReadDB test still went through the ReadDB server and the BAM and WIG files were read by the httpd process. These tests indicate that ReadDB's fast queries compared to remote BAM are not the result of BAM queries passing more information across the network; the ReadDB queries through the ReadDB server are faster than either BAM through httpd or BAM read directly from local disk.

P Alexander Rolfe, et al. BMC Bioinformatics. 2011;12:278-278.
3.
Figure 4

Figure 4. From: ReadDB Provides Efficient Storage for Mapped Short Reads.

Performance comparison across a high-latency network. Comparison of runtimes across a residential broadband connection. The query times correlate with the amount of data transfered over this slower connection-the histogrammed BigWig results are fastest, followed by the per-read ReadDB results (readdb) followed by the BAM results. The server-side histograms with ReadDB (readdb.histogram) demonstrate the best performance as they combine histogrammed data transmission with the fast ReadDB server. Each method was used to query n randomly chosen 1 kb regions; only a single replicate was run. The readdb. histogram test used server-generated histograms such that the data passed across the network is similar to the BigWig format rather than the per-read format in the bamlocal and readdb tests.

P Alexander Rolfe, et al. BMC Bioinformatics. 2011;12:278-278.
4.
Figure 2

Figure 2. From: ReadDB Provides Efficient Storage for Mapped Short Reads.

ReadDB performance compared to other formats. ReadDB's performance nearly matches that of local BAM or BigWig files on the medium and large datasets. Each method was used to query n randomly chosen regions (of size 1 kb, 10 kb, or 100 kb) and the resulting average query time is shown in milliseconds. The client and server machines were connected by gigabit ethernet. ReadDB and the BAM tests retrieved individual hits; BigWig retrieved a histogram in 10 bp bins as the BigWig format cannot store individual read positions. bamlocal and wiglocal queried files on local disks rather than on a remote server.

P Alexander Rolfe, et al. BMC Bioinformatics. 2011;12:278-278.

Supplemental Content

Recent activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...
Write to the Help Desk