Frequently Asked Questions
Q: Why is the system named "Rockhopper"?
Q: What if my RNA-seq data is not in one of the formats supported by Rockhopper? Q: What if I did or did not preserve strand information in my RNA-seq experiment? Q: What if my experiment used size-selected samples? Q: Does Rockhopper support single end reads? Does Rockhopper support paired-end reads? Q: Does Rockhopper support older encodings of Phred quality scores? Q: How long does Rockhopper take to execute? Why is it faster during some executions than others? Can it be parallelized? Q: What if my reads do not align as expected? Q: Why does Rockhopper not work, stall or abort execution? Q: Can I use Rockhopper to analyze RNA-seq data from one or more organisms whose genomes have not yet been sequenced? Q: What do the expression values for each transcript correspond to? Q: What are the q-values reported by Rockhopper? What is the difference between q-values and p-values? Q: Is Rockhopper compatible with other RNA-seq analysis tools?Answers to Frequently Asked Questions
| Why is the system named "Rockhopper"? |
Rockhopper is a type of penguin. The name has no significance.
| What if my RNA-seq data are not in one of the formats supported by Rockhopper? |
Rockhopper accepts files in the most commonly used formats: FASTQ or QSEQ or FASTA or SAM or BAM. FASTQ and QSEQ and FASTA files optionally may be gzipped. Future versions of Rockhopper may include support for additional file formats.
| What if I did or did not preserve strand information in my RNA-seq experiment? |
Rockhopper supports RNA-seq experiments both when strand information is preserved and when strand information is not preserved. By default, Rockhopper assumes strand information is preserved, but this can by modified in the Parameter Settings window.
| What if my experiment used size-selected samples? |
Rockhopper supports analysis of data from size-selected RNA-seq experiments. If too many small transcripts are identified by Rockhopper, the expression threshold for identifying small transcripts can be adjusted in the Parameter Settings window.
| Does Rockhopper support single end reads? Does Rockhopper support paired-end reads? |
Yes. Rockhopper supports both single end reads and paired-end (mate-pair) reads.
| Does Rockhopper support older encodings of Phred quality scores? |
Yes. Different Illumina formats use different encodings for Phred quality scores. Rockhopper automatically detects the encoding and decodes to the appropriate Phred quality score.
| How long does Rockhopper take to execute? Why is it faster during some executions than others? Can it be parallelized? |
The first time that Rockhopper aligns sequencing reads to a genome, it typically takes several minutes, depending on the number of reads, the length of the reads, the size of the genome, and the speed and number of processors on your computer. Rockhopper automatically detects the number of processors on the computer and uses all available processors. As a crude guide, Rockhopper takes about a minute per million reads per processor. After the first time that reads are aligned to a genome, Rockhopper caches the results so that future executions of Rockhopper using the same reads and parameters should execute in a matter of seconds.
| What if my reads do not align as expected? |
The most common reason that reads align predominantly antisense rather than sense or for some other reason do not align as expected is that the orientation of the reads is not properly set. The orientation of the reads can be adjusted in the Parameter Settings window using the Reverse complement reads option for single-end reads and using the Orientation of mate-pair reads option for paired-end reads.
| Why does Rockhopper not work, stall or abort execution? |
There are two common reasons why Rockhopper may not execute
as expected. If Rockhopper cannot create files on your computer, then it
will fail. For example, on a Mac, if the Rockhopper application
is not moved to the Desktop or some other location where it
can write files, then it may fail. The other common reason
why Rockhopper may not execute as expected
is a lack of memory (RAM) available to Rockhopper.
The Windows and Mac downloadable verions of Rockhopper, by
default, allocate 1.2 gigabytes of memory to Rockhopper, which
is sufficient memory for most applications. However,
when data from a large number of RNA-seq
experiments is being analyzed, Rockhopper may require
more memory. In this case, the downloadable JAR version of
Rockhopper should be used on a 64-bit machine running
64-bit Java. For example, to allocate 4 gigabytes
of memory to Rockhopper, the JAR file can be executed as
follows:
java -Xmx4000m -jar Rockhopper.jar
| Can I use Rockhopper to analyze RNA-seq data from one or more organisms whose genomes have not yet been sequenced? |
Yes. Rockhopper supports de novo transcriptome assembly when no reference genome is available.
| What do the expression values for each transcript correspond to? |
Expression values reported by Rockhopper for each transcript in each condition are similar to RPKM (reads per kilobase per million mapped reads) values. However, RPKM values are generally normalized by the total mapped reads, whereas the expression values reported by Rockhopper are normalized by the upper quartile of gene expression, which is a more robust normalizer.
| What are the q-values reported by Rockhopper? What is the difference between q-values and p-values? |
To determine whether a transcript shows differential expression in data from two conditions, Rockhopper performs a statistical test for the null hypothesis, which is that the expression of the transcript in the two conditions is the same. Using a Negative Binomial distribution as its statistical model, Rockhopper computes a p-value indicating the probability of observing the transcript's expression values by chance. Because multiple tests are being performed across the set of transcripts, Rockhopper reports q-values, which are adjusted p-values, that control the false discovery rate using the Benjamini-Hochberg procedure. The p-values (and subsequently q-values) reported by Rockhopper are influenced by many factors, including the number of replicate experiments in each condition, the variance of a transcript's expression across replicates, and the size of the transcript.
| Is Rockhopper compatible with other RNA-seq analysis tools? |
Rockhopper can read in files in SAM and BAM format, commonly used by RNA-seq analysis tools. Rockhopper can also output files in SAM format, which can then be used by other tools.