High Throughput Genome Builds

From genomewiki
Jump to: navigation, search

This is a transcript of whiteboard notes from the 5/30/06 Genecats meeting.

Input Requirements

AGP File, Sequence

Output Tracks

Gap. Gold, ChromInfo, GC5

Then Repeat Masking/Window Masker and trfBig

Produces rmsk/windowMask and simpleRepeats track

Early Tracks

Genscan, Genbank Scripts, CpG Islands

Map Closest protein set (human mostly)

Same Species LiftOvers

Comparative Genomics

human and mouse

Candidate Genomes

(A)platypus(new + update), (A)frog update, (A)Stickleback(fish)(new + update), (A)Medaka(fish), (B)rhesus update, (B)chicken update, (A)elephant, (B)rabbit(new + update), (A)pig(new + update), (B)cat(new + update), (A)fugu update, (B)chimp update, (C)mouse, (C)human

Key: (A) = Abbreviated browser only (B) = More expanded browser (C) = Full browser

Tools

  • xxxToAgp
  • agpToDb builds to to RepeatMasking, include data staging ? SAN, scratch, iscratch

    verifies AGP & Sequence, sets up unmasked sequence into 500K chunks for RM, stores sequence in standard place.

  • masker with or without repeat library

    Work towards windowMask automation

  • Masked sequence stager
  • make goldenPath downloads
  • makeXxxTrack

    Work towards track by track automation

  • doBlastzChainNet (how about saving run-time parameters in a metadata DB table to create README/.html information ?)

    chain/net .html file $alignSettings keyword and library function to make alignment parameter table

  • Conservation (On all browsers ?)

Bugs

.html pages for tracks

QA Time

  • Downloads
  • Metadata, hgCentral, defaultDb, dbDb, all.joiner, READMEs
  • liftOvers (between same species)
  • HTML Pages
  • Links and Clickthroughs


Metadata blastz parameter table

OrganismAssemblymatrixgapsabridged repeats...etc...
mousemm8mediummediumyes...etc...
humanhg19looselooseno...etc...