TITLE:
SNP detection by Applied Biosystems SOLiD sequencing of HapMap individual NA19240
AUTHORS: AUTHORS: McLauglin S, Peckham H, Hyland,F.C., Scafe C., Costa G., Muzny D., Gibbs R., De La Vega F.M., McKernan K.
YEAR: 2008
STATUS: 1
||
TYPE: METHOD
HANDLE: 1000GENOMES
ID: SOLiD.Corona.v2
METHOD_CLASS: Sequence
SEQ_BOTH_STRANDS: YES
TEMPLATE_TYPE: DIPLOID
MULT_PCR_AMPLIFICATION: NA
MULT_CLONES_TESTED: NO
METHOD: 

Summary:

The YRI 1000 Genomes sample NA19240 was sequenced with the SOLiD System at Baylor College of Medicine HGSC and at Applied Biosystems. The sample was sequenced to 29x.

Mapping and SNP Calling:

Mapping, pairing, and SOLiD SNP calling was performed with v2 of the SOLiD Corona SNP caller (source 
code can be downloaded at www.solidsoftwaretools.com).  
SOLiD Corona SNP caller is a heuristic rules-based method for calling SNPs, based on removing sequencing error while retaining evidence of a genetic variant using SOLiD colorspace rules. The method incorporates information about coverage, 
number of observations of the reference allele and the alternate allele, quality values of the reads, resulting in a 
score reflecting the characteristics of the reads providing evidence of a SNP.

Filtering:

The SNPs list was filtered to eliminate candidiate false positives SNPs as follows (about 4% of total SNPs were removed after filtering).

Novel SNPs within 10bp of any dbSNP entry	 	 	   	   93,674
Within 10bp of a SOLiD Indel   	 	 	 	 	   21,600
Invalid Double SOLiD SNPs      	 	 	 	 	   21,465
SNPs with >75x coverage        	 	 	 	 	   17,055
Heterozygotes in SOLiD CNVs with increased copy number  	   15,112

* Notes:  we identified all SNPs within 10bp of any dbSNP entry (SNPs and InDels) but we removed 
only those that were novel.  For the other 
filters, we removed both dbSNP and novel SNPs.
SOLiD Indels means those small InDels discovered in this sample using these 
SOLiD sequencing runs, and similarly with SOLiD CNVs.

Further filtering removed all SNPs which were not also detected in this family 
using the Sanger Centre trio calling method applied to Illumina high-throughput 
sequencing data averaging 12.6, 15.1 and 17 x coverage for individuals NA19238, 
NA19239 and NA19240 (mother, father and daughter) respectively.


Validation:

The validation was done by Baylor College of Medicine HGSC for only novel SNPs, ie those not in dbSNP129. 
Validation was done by sequencing, using primers already in Baylor freezers 
(so biased towards SNPs in Encode regions but not in dbSNP, and towards SNPs 
in exons but not in dbSNP, since these were over-represented in the freezers).  
AB provided a SNP list, Baylor randomly selected from this list SNPs covered 
by primers in their freezer. Validation was performed by Sanger sequencing. 
A major advantage of validation by Sanger sequencing is that the primers were
specificially designed to amplify regions that may contain repeats, homologous 
genes, etc;  that is, regions hard to access by genotyping and by arrays, and 
so more difficult regions than the 'HapMappable' genome. These should be 
regions representative of sequencing results on the whole genome (excluding 
highly repetitive regions such as telomeres, centromeres, Y chromosome, etc) 
and not just the 'easier to sequence and map' regions represented by arrays 
and HapMap.
 
In the two summary data files, the interpretation of the columns is as follows:

1. Number of reads covering the position with 2 color space calls (i.e., reads 
only partially covering the base with the last color space call of the read aren't 
counted)

2. Reference base at this position

3. Consensus base at this position (IUPAC codes are used for heterozygotes)

4. Score* of the reads of the Reference allele

5. Confidence* of the reads of the Reference allele

6. Number of single tag (from fragment runs) reads containing the reference 
allele / number of unique start positions of reads containing the reference allele

7. Number of mate pair tags (from paired-end runs) reads containing the 
reference allele / number of unique start positions of reads containing 
the reference allele

8. Score* of the reads of the non-reference allele

9. Confidence* of the reads of the non-reference allele

10. Number of single tag (from fragment runs) reads containing the non-reference 
allele / number of unique start positions of reads containing the non-reference allele

11. Number of mate pair tags (from paired-end runs) reads containing the 
non-reference allele / number of unique start positions of reads containing 
the non-reference allele

12. Base space coordinate of the position (1-based)

13. Chromosome Number (where 23=X and 24=Y)          

14. Chromosome name                                        

15. rs_id if in dbSNP.                                        

Score* = the weighted score assigned to the reference or consensus base, 
based on all the reads that cover it with 2 color space calls (the weighted 
scores for each of the 16 types of dibase combinations sum to 1)

Confidence* = the average confidence of each read that was used to generate 
the weighted score of the reference or consensus base (aka Valid Score)

||
TYPE: POPULATION
HANDLE: 1000GENOMES
ID: 1000GENOMES_pilot_2_YRI_child
POP_CLASS:  west africa
POPULATION: One individual, NA19240, from the HapMap YRI sample
||
TYPE: SNPASSAY
HANDLE: 1000GENOMES
BATCH: 1000GENOMES-NA19240-2008-12-12
MOLTYPE: Genomic
METHOD: SOLiD.Corona.v2
SAMPLESIZE: 2 chromosomes
ORGANISM: Homo sapiens
COMMENT: initial data release
||