Manhattan

[diegohernansanchez]

Hi,
Not really an Issue. I was wondering if you have at hand a solution to articulate Manhattan plot data on pyGenomeTracks
Thanks
diego

[lldelisle]

Hi,
Could you describe a bit more what you mean by Manhattan plot and what is your input data?
Thanks,
Lucille

[diegohernansanchez]

Hi Lucille,
Thanks for prompt response.
Sure, a Manhattan plot depicts association-mapping data. This is a statistical correlation between a phenotype and detected genetic polymorphisms at population level. If a particular polymorphism (say a single nucleotide polymorphism -SNP-, which is a base variant in a particular chromosomal coordinate) occurs in a fraction of the population and is significantly correlated with a phenotype, then this polymorphism is given a lower p-value. Otherwise, non-correlated SNPs present a higher p-value. Hence, a Manhattan plot depicts -log(p-values) along chromosomal coordinates (see an example in https://en.wikipedia.org/wiki/Manhattan_plot).
It would be very useful that pyGenomeTracks could handle this type of data over a genome browser, as genetic causality could be supported by comparison with additional genome-wide data such as RNA-seq, etc. The input data is simpler than a BED4, with just a single chromosomal coordinate instead of start-end. It would be a tab-delimited .txt file with columns: chromosome | position | -log(p-values). It is just needed to depict a dot in that position, with certain size and color, the latter usually depending on the chromosome number.
Thanks for your time. Cheers
diego

[lldelisle]

Hi,

• For me, if you manage to convert your input file to bedgraph you can use the type = points or type = points:0.5 where you specify the point size. For example, if your input file has only 3 columns (chr | pos | -log(pval)) you can use awk to convert it:
  awk -v OFS="\t" '{print $1,$2,$2+1,$3}' input_3col.txt > output.bedgraph.
• For the moment it is not possible to plot multiple chromosomes on the same x-axis (you need to do one plot per chromosome).
• We do not support to automatically change color per chromosome so you would need to do a ini file per chromosome.
  Cheers,
  Lucille