Methodology

This page documents Round 1 of converting soybean genomic data so that it can be displayed in GBrowse. Some 4,231 features were loaded into GBrowse: 3,367 Loci, 769 Clones, 75 Contigs, and 20 MLG's.

Distinguishing characteristics of Round 1:

• This was a first effort in converting existing data into GFF format so that it could be accessed by GBrowse.
• Many small steps were taken to analyze and convert the data. In many cases, particular steps were inefficient.
• Many features were placed prior to the contigs being placed, and then had to be moved.
• The result had to be visually inspected with many manual adjustments being made.
• The data which was anchored to an MLG did not properly show the makeup of the contigs.
• Ontology: The following features were placed.
  • Σ Loci
  • Σβ₁ Clones and γ Clones
  • γ Contigs

A centimorgan-to-base conversion factor of 441,824.931130786 was used. An explanation as to how this was determined is in Step 1, Phase 2 A band-to-base conversion factor of 2,887.52588816 was used. An explanation as to how this was determined is in Step 3, Phase 1.

In some cases, the same marker was the anchor for two or more contigs. In these situations, the contigs were moved to be side-by-side.

The methodology of Round 1 follows below. To start at the beginning, read from the bottom up..

When the features were converted from the genetic map to the physical map, many of them inappropriately appeared to land in the same positions, which is impossible. Staggering their positions resulted in Step 5, which continued from the initial setup described further below.

(Note that in the process of staggering the loci, it was noticed that A724_3 and A510_2 each appeared on two different MLG's.)

Step 5 had two dead-end phases in its development. In Step 5, Phase 1, the locations were staggered in 1,000-base increments. This solved the problem of the loci overlapping, but clones and contigs still had inappropriate overlaps. Step 5, Phase 2 adjusted these increments to 100,000 bases. However, the clones and contigs would have to be readjusted and then the clones and loci readjusted. It was easier to stagger the contigs, first, intending to readjust the clones and loci. Step 5, Phase 3, did this. Step 5, Phase 4, adjusted the locations of the anchor clones which were associated with the contigs. Step 5, Phase 5, adjusted the locations of the anchor loci. Step 5, Phase 6, adjusted the locations of the corresponding non-anchor clones which only had linkages to one locus each. Step 5, Phase 7, manually adjusted the locations of six clones which had linkages to two loci each. (As a result, two other loci also had to be moved.) Remaining loci which were still stacked were staggered in Step 5, Phase 8. The corresponding clones were moved in Step 5, Phase 9.

The methodology of adjusting the features with conflicting locations follows..

• Step 5, Phase 10: soybean.gff Version 0.24, 01.soybean.conf Version 0.16, July 21, 2003
  • Some 199 features were manually adjusted because of conflicts .
     
     
• Step 5, Phase 9: soybean.gff Version 0.23, 01.soybean.conf Version 0.16, July 8, 2003
  • Some 495 clones were adjusted to correspond with the loci moves in the previous step. Fifteen other features were manually adjusted..
     
     
• Step 5, Phase 8: soybean.gff Version 0.22, 01.soybean.conf Version 0.16, July 7, 2003
  • Some 491 loci which were still inappropriately stacked were staggered. Eleven more were manually adjusted.
     
     
• Step 5, Phase 7: soybean.gff Version 0.21, 01.soybean.conf Version 0.16, July 7, 2003
  • Six non-anchor clones with linkages to two loci each were moved to correspond with their anchor loci.
     
     
• Step 5, Phase 6: soybean.gff Version 0.20, 01.soybean.conf Version 0.15, July 6, 2003
  • Some 142 non-anchor clones with linkages to one locus each were moved to correspond with their anchor loci.
     
     
• Step 5, Phase 5: soybean.gff Version 0.19, 01.soybean.conf Version 0.15, July 5, 2003
  • Loci locations were moved to correspond with the anchor clones. This was done exactly the same way as in Step 4, Phase 2 Loci anchoring clones were labeled with the word "Anchor" in the GFF notes field so that they would not be moved, again.
     
     
• Step 5, Phase 4: soybean.gff Version 0.18, 01.soybean.conf Version 0.15, July 5, 2003
  locations were moved to correspond with their contigs. This was done exactly the same way as in Step 4, Phase 1 Clones anchoring contigs were labeled with the word "Anchor" in the GFF notes field so that they would not be moved, again.
   
   
• Step 5, Phase 3: soybean.gff Version 0.17, 01.soybean.conf Version 0.15, July 5, 2003
  • The loci were relocated to their original positions (temporarily) and 19 contigs were staggered, instead.
     
     
• Step 5, Phase 2: soybean.gff Version 0.16, 01.soybean.conf Version 0.15, July 4, 2003
  • (A dead-end phase of development.) Some 518 loci locations were re-staggered in 100,000-base increments (instead of 1,000-base increments) so that their positions did not overlap with each other. QTL starting and ending location increments were increased to 7.5 cM to represent ranges of their possible locations (rather than their lengths).
     
     
• Step 5, Phase 1: soybean.gff Version 0.14, 01.soybean.conf Version 0.15, July 2, 2003
  • (A dead-end phase of development.) Some 497 loci locations were staggered in 1,000-base increments so that their positions did not overlap with each other.
     
     

The determinations of the initial locations of features resulted in an interesting loop: The contig locations depend upon the clone locations; but the clone locations depend upon the Loci locations; but the Loci locations depend upon the contig locations. Therefore the following initial plan was followed:

The initial plan:

• Step 1: Load QTL's and Loci.
• Step 2: Load clones based on the locations of the QTL's and Loci.
• Step 3: Load contigs based on the locations of the clones.
• Step 4: Adjust the locations of the markers based on their locations in the contigs.

The methodology of entering the initial soybean data into GBrowse follows. The more recent changes are at the top. To start from the beginning, read from the bottom up.

• Step 4, Phase 2: soybean.gff Version 0.13, 01.soybean.conf Version 0.15, June 20, 2003
  • Some 62 loci locations were slightly changed based on their relationships to the contigs whose locations were just changed.
     
     
• Step 4, Phase 1: soybean.gff Version 0.12, 01.soybean.conf Version 0.15, June 20, 2003
  • Some 116 clone locations were slightly changed based on their comparative locations in contigs.
     
     
• Step 3, Phase 1: soybean.gff Version 0.11, 01.soybean.conf Version 0.15, June 18, 2003
  • Some 75 contigs were added to the GBrowse database.
     
     
• Step 2, Phase 2: soybean.gff Version 0.10, 01.soybean.conf Version 0.14, June 17, 2003
  • Eleven clones matching exactly two markers were loaded into GBrowse.
     
     
• Step 2, Phase 1: soybean.gff Version 0.09, 01.soybean.conf Version 0.13, June 17, 2003
  • Some 758 clones matching exactly one marker were loaded into GBrowse.
     
     
• Step 1, Phase 2: soybean.gff Version 0.08, 01.soybean.conf Version 0.12, June 14, 2003
  • Three changes were made to the database: (1) the lengths of the QTL's were changed as described within; (2) the locations of the loci were slightly adjusted as described within; and, (3) The total base size was changed from 1,100 million bases to 1,115 million basess ^{Arumuganathan}. Phase 2 also improved the Perl scripts so that changes like these are easier to make.
     
     
• Step 1, Phase 1: soybean.gff Version 0.07, 01.soybean.conf Version 0.12, June 11, 2003
  • Twenty MLG's and 3,367 QTL's and Loci were loaded into GBrowse.
     
     
• Step 1, Phase 0: soybean.gff versions 0.06 and below, 01.soybean.conf version 0.11 and below, pre-June 11, 2003
  • In these tries the data output was either incomplete or incorrect, so they were not documented here.