1. This project described and tested a method of analysis, based on a linear mixed model, that accounts for the correlation structure of recombinant inbred intercross (RIX) populations. For detailed information see Tsaih et al  Mammalian Genome  (2005) 16: 344-355; 


2. The data files of the project was produced by the Map Manager QTX program. They include genotype file (for CxB RI lines only), phenotype file (includes both RI and RIX line) and genetic map file. Genotype data of CXB RI strains and phenotype data of CXB RI strains and RIX hybrids were provided by the Williams Lab in Univeristy of Tennessee. Most of the data can be found at http://www.nervenet.org. The map file was expanded by the authors prior to QTL analysis, and it has been updated with Build37 Cox genetic map (B37), the new map file is saved as "Tsaih2005_B37_map.txt". The updated genotype file ("Tsaih2005_B37_geno_RI.txt"), phenotype file ("Tsaih2005_B37_pheno_RIX.csv") and map file can be combined by read.cross function in Rqtl by setting the format parameter as "mm". For the details of the update, please see the information below; 


3. This mapping population consists of 78 nonreciprocal CxB RIX hybrids, a complete half-diallel cross of the 13 parental CxB RI strains. The F1 intercrosses between pairs of CXB RI strains were set up to generate all 78 nonreciprocal matings by crossing low-numbered female strains by higher-numbered male strains. This simple low-by-high breeding scheme results in a systematic bias: CXB1 is always used as a maternal strain and CXB13 is always used as a paternal strain. For detailed protocols on RIX mice breeding and sample collections see http://www.nervenet.org  or Williams et al., 2001; 


4. The CxB RIX lines didn't have genotype data. CxB RI genotype data was used to impute the genotypes of 78 nonreciprocal CxB RIX hybrids in this paper. The imputed RIX genotypes were then used to simulate new phenotypes to test the proposed method. All the steps were accomplished by the R codes made by authors and we don't have the imputed genotypes of RIX lines; 


5. Generate a description file of the project, it includes the general information, definitions of the phenotypes, description of the genotypes and information of the missing markers;


6. For the genotypes of this cross: (cM=centimorgan, bp=basepair)
(1) Change the symbol of heterozygous allele from "B" to "H", missing genotypes from "" (original data) to "-" and save the changes in "Tsaih2005_B37_geno_RI.txt" (B37 data file); 

(2) Marker "D1Nds4" is unmapped in Cox map, primer sequences found in MGI, and their locations identified using primer-BLAST in NCBI followed by UCSC In-Silico PCR. The only issue about this marker is the primer prodouct size in MGI is 107bp, the product length from NCBI primer-BLAST is 160bp, there is 53 bp difference between them. I used result from NCBI to get the current basepair position, then convert it to cM through Map Converter tool. Marker "D17Mit1" has a same problem, the length difference between MGI primer and product of NCBI is 52 bp, as same as before, I used result from NCBI to get the current basepair of this marker (this type of markers is highlighted with light green color in description file);

(3) The bp position of marker "D5Mit267" is missing in both Cox map and MGI. Primer sequences found in MGI, then use primer-BLAST in NCBI to get primers, but the UCSC In-Silico PCR doesn't have any matched results for these primers so we could not identify the B37 bp position of this marker currently (this type of markers is highlighted with light blue color in description file);

(4) Unable to find cM position of marker "D8Mit181" from Cox map. The current symbol of this marker in MGI is "D5Mit181" and it is mapped on chr 5, not chr 8. The B37 data file uses the result from MGI. Its bp position assigned based on MGI, then convert it to cM through Map Converter tool (this type of markers is highlighted with teal color in description file);

(5) The bp position of marker "D12Mit14", "D14Mit156" and "D14Mit166" are missing in both Cox map and MGI. Primer sequences found in MGI, but no target templates were found in primer-BLAST in NCBI. We could not identify the B37 bp positions of these markers currently (this type of markers is highlighted with dark yellow color in description file);

(6) The bp position of marker "D10Mit205" and "D14Mit75" are missing in both Cox map and MGI. No primer sequences are available for them from MGI. We could not identify the B37 bp positions for these markers currently (this type of markers is highlighted with pale blue color in description file);

(7) Marker "D14Mit168" is unmapped in Cox genetic map. The MGI shows this marker should be on chromosome 14, but it doesn't have any cM, basepair or primer information in database. So I ignored the name of the marker and still think the marker was on chromosome 19 in this project only. Use the interpolating algorithm in Rqtl to calculate its B37 cM position, the bp information is unavailable (this type of markers is highlighted with orange color in description file);

(8) Marker "D6Mit198", "D9Mit146" and "D19Mit93" are unmapped in Cox genetic map, the primer sequences could not be found in MGI, either. Use interpolating algorithm in Rqtl to calculate the B37 cM position, the bp information of them are unavailable (this type of markers is highlighted with lavender color in description file);

(9) Ten markers miss bp positions in both Cox map and MGI. Primer sequences found in MGI, but their locations don't match the result from primer-BLAST in NCBI. We could not identify the B37 bp position of this marker currently (this type of markers is highlighted with rose color in description file and please see description file for details);

(10) Twelve markers miss bp positions in both Cox map and MGI. Primer sequences found in MGI, and their locations identified using primer-BLAST in NCBI followed by UCSC In-Silico PCR (this type of markers is highlighted with lime color in description file and please see description file for details);

(11) Six markers are unmapped in Cox genetic map. Their bp positions assigned based on MGI, then convert them into cM through Map Converter tool (this type of markers is highlighted with light yellow color in description file and please see description file for details); 

(12) 34 markers miss bp positions in Cox genetic map, assign the positions from current MGI database. This type of markers are highlighted with gold color and please see description file for details; 

(13) Some markers are at the same position (cM) in B37 Cox map:


7. For the phenotypes of the cross:
(1) The brain weight data of CxB RI strains and RIX hybrids were from the Williams Lab in Univeristy of Tennessee. 
For detailed protocols see http://www.nervenet.org/main/databases.html  or Williams et al., 2001.   

(2) The information of intervention (diet type) and mice testing age are not found in paper (highlighted in turquoise color);

(3) Could not find the definitions of trait "mean_eye", "mean_cere", "s_body", "s_brain", "s_eye", "s_age" and "s_cere" in paper, they are highlighted in turquoise color in description file;

(4) Change the "F" letter to "0", "M" to "1" in trait "Sex", change the missing values from "" to "-" and save the changes in "Tsaih2005_B37_pheno_RIX.csv" file;


8. Except the files mentioned above, I also saved the description file, original data and map files, update data and B37 map files, data process and list of missing markers together as a big excel file ("Data_Description_RIX_Tsaih2005.xlsx") (1st curating work is finished on 7/9/2009, files are updated on 2/23/2011).


9. Information were updated on 5/25/2012:
Change the name of the map (from "Shifman" to "Cox" map) in description and readme file, add a map reference in description file.