Thursday, October 31, 2013

v26 (3 of 3): smoothing curves

A minor improvement has been made that allows you to smooth the graph of a quantitative track using window mean values.

An un-smoothed numerical track might look like this:


To smooth the track, right click on the track graph and choose "Configure". Check the "Apply smoothing window" checkbox, and adjust window size:


The graph updates after applying the 9-pixel smoothing window to the red track:





v26 (2 of 3): RepeatMasker tracks

A minor update to RepeatMasker tracks has been made, which is the ability to grade repetitive elements by their Smith-Waterman alignment scores.

The updated RepeatMasker tracks are available for following genomes: hg18, hg19, mm9, mm10, rn4, danRer7, dm3. All the RepeatMasker track data were downloaded from UCSC Genome Browser.

Go to browser and show human hg19 genome. At Tracks > Genomic annotation tracks > RepeatMasker, you can see a long list of tracks. Each track is one "class" of repeats, click the arrow following the track to see tracks of families from this class:


Click track L1 (LINE) to display it, which is the "L1" family from "LINE" class:



The L1 elements in above screenshot show various shades of color, which are determined by the greatest SW score of all L1 elements in the view range. Click on an element to see details:


Right click on the L1 track and choose "Configure". You can find options about what kind of scores can be used to grade the elements, or you can choose to disable the score and all items will become fully visible:


v26 (1 of 3): SNP track


SNP information is a long-desired feature and we've finally got it on our browser. Credit goes to UCSC Genome Browser for making SNP information nicely curated and easily accessible.

In version 26 only the human hg19 genome has its SNP track (dbSNP release 137). We are working to add SNP tracks for all applicable genomes as well as enabling SNP custom tracks.

Click to see the SNP track.


Display hg19 SNP track

Go to the browser website and display human hg19 genome. Click "Tracks" > "Genomic annotation tracks", and find SNP track in "Variation" category:


Click the SNP track item to show the track on the browser:


By default the SNP track is shown in "full" mode, with the "rs..." name printed on the left of SNP location. SNPs are colored differently according to their classification:


  •       single nucleotide variation
  •       insertion/deletion
  •       heterozygous variation
  •       microsatellite
  •       text name but not sequence
  •       a cluster of multiple classes
  •       multiple nucleotide polymorphism
  •       insertion
  •       deletion
Click on a SNP to see details about it:


Erratum: rs78276647 should be on "forward" strand


You can use the "link" on the top-right of the tooltip to see original record of this SNP in dbSNP website.


Search for SNPs
To search for SNPs, you need to have SNP track displayed first.

Right click on SNP track and find the search box:


Enter SNP name in the box and press Enter to find the hit:


Click the hit to show it in browser. The hit SNP will be highlighted:



Get v26 source code here, or here.

Wednesday, October 9, 2013

v25: new display methods for methylC-seq and RRBS experiments

Version 25 features a nice way to display CpG methylation data, and the track background color configuration.


CpG methylation track
Tracks like those generated by methylC-seq (bisulfite sequencing), and Reduced Representation Bisulfite sequencing (RRBS) assays are single-base resolution. There won't be any data unless there's a CpG site. Traditionally such tracks are displayed as numerical track with a number ranging from 0-1 for all bases. That leads to a confusion of "0 value (no methylation)" and "lack of data (not a CpG)".

Now a new display method is available to address this issue:




This example shows two methylC-seq tracks in human hg19. Methylation levels over CpG sites are marked out by red/blue bars, and the "counter-value" is marked out with gray. A long gray bar means the CpG site has very low methylation level. This way you can tell a unmethylated CpG site from those blank regions with no CpG sites.

This rendering method is suitable for numerical data with incomplete coverage of the entire genome. In the other words, if your track hasn't got a value for every bp in the genome, you should use this display method.

To use this display method, convert your track into bedGraph format.

To turn on the effect, add this attribute in the track object in the datahub:

barplot_bg:"#cccccc",

Or in the embedding code, add this:

barplotbackground:"#cccccc",




Track background
Now it's possible to apply background color to tracks. Simply check the "use background color" option in Configure menu:



To apply background color in datahub or embedded browser, add this in track object:

backgroundcolor:"#cccccc",