Detailed steps in MARS workflow which are linked to GDMS
1. The breeder has already selected the parental lines and made the population
2. Set up the project (define the parents and the type of population e.g., F3:5, any need for diagnostic loci to be included)
3. Breeder queries GDMS for the parental lines. Query requires:
a. Fingerprint data (if lines have been fingerprinted. If not require parental screening with marker set) to identify only polymorphic markers between parents
- For all markers that have been fingerprinted on both lines bring back only those that are polymorphic.
- System produces an interactive map of polymorphic markers (flag diagnostic loci)
- Breeder needs to be able to select markers on chromosome, position, desired density, type.
- Genotyping order created based on selected markers and individuals in population plus both parents used in initial cross.
- Results from external genotyping provider are stored in GDMS to be used in subsequent analyses.
b. Genotypic data is used in analytical pipeline to identify marker trait associations (QTL)
- Breeding View/RAP to access genotypic data for population and parents from GDMS
- Breeding View/RAP will merge phenotypic and genotypic datasets for QTL analysis
4. User selected QTL stored in GDMS. Mechanism for automatic transfer of information to GDMS.
5. Selected QTL data for the project retrieved from GDMS to be loaded in OptiMAS.
6. Selected progenies identified.
7. Genotype x individual plants from each progeny using the QTL loci and select 1 best plant from each progeny to recombine.
8. Go through successive recombination cycles.
- Genotypic information can be handled locally (outside of GDMS during the recombination cycles).
9. Final individuals selected which will be evaluated in the field.
Kevin: From the start it seems that it is expected that GDMS already holds data, is this correct? So the GDMS database to be installed on workstations will be populated by available data?
Mark: My expectation is that GDMS will be populated with data and is it something that our colleagues, Chunlin and team at ICRISAT are working on. Certainly the fingerprint data. Then we need to consider all Marker Trait Association data (e.g., QTL data) which may already be available or will come later when the analysis is done.
Kevin: How will Breeding View/RAP access genotypic data from GDMS? Will it directly access the database or will it read output files generated by GDMS? Is this functionality included in the June deadline?
Mark: Directly access the database. Yes there is some functionality included in the June deadline.
Kevin: For "User selected QTL stored in GDMS. Mechanism for automatic transfer of information to GDMS." the information to be stored comes from Breeding View/ RAP? Would this info be on a file which can then be used as input to GDMS?
Now we can automate storing of information from a file to GDMS if Trushar's team can implement a method which can perform the process. Currently in the GDMS application, this insertion of data requires user interaction for selecting the file. The code to insert the info to the database is already there, what is needed is some method which can accept as parameter the location of the input file and this method can be called on the command prompt. Like for example, issuing a command: "java -jar gdms.jar input-file.xls".
Mark: Yes. The final part of the QTL analysis is the user selecting (using various criteria) those QTL for the project that he/she wants to store for later use by OptiMAS. Expectation is that the storage with be done by GDMS and then OptiMAS will access this data for the next stage of the analysis
Kevin: Are the output files which GDMS generates accepted by OptiMAS? If not then we should define how to convert the output from GDMS to make it acceptable by OptiMAS.
Mark: I do not know what the output files that GDMS generates look like. However I have circulated the relevant input files widely so that we should be well aware what the inputs to OptiMAS look like (both QTL and marker data).