discrepancies in allele frequencies between 1000 genomes and 5400 exomes

AJERYC · 08-22-2012, 05:29 AM

Using ANNOVAR summary routine I have got a exome_summary.csv file that contains different columns with 1000 genomes and 5400 exomes frequencies with some empty cells. I asume that empty cells correspond to SNPs that were not found in any of these databases. However there are cases where you get frequencies either in 1000 genomes or 5400 exomes but not in the other database. I would understand it if the alllele frequency in the alternate database would be very low, but there are cases where this frequency can be up to 80%. Does anyone know what is the reason for these big discrepancies between these databases?

etal · 10-31-2013, 11:01 AM

The 1000G and 5400ESP projects used different pipelines for sequencing, read mapping and and variant calling, so at least some of the major discrepancies between the listed allele frequencies are due to different quirks in their respective pipelines. For example, each pipeline might consistently map ambiguous reads to a different site in a repetitive or low-complexity region, giving the variant calls in that region a high frequency in one database but not the other.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Allele frequency per ethnic group in 1000 genome project	izhar	Bioinformatics	1	05-20-2012 03:25 AM
Need help identifying SNPs and allele frequencies	dkenned1	Genomic Resequencing	1	11-06-2011 03:02 PM
Need help identifying SNPs and allele frequencies	dkenned1	Bioinformatics	1	11-06-2011 01:17 PM
Get allele frequencies for specific coordinates from a .bam file	mehc	Bioinformatics	1	10-28-2011 01:05 PM
bam files of exomes from 1000 genomes	jorjial	Bioinformatics	3	06-06-2011 04:16 AM

08-22-2012, 05:29 AM	#1
AJERYC Member Location: Spain Join Date: Jan 2012 Posts: 26	discrepancies in allele frequencies between 1000 genomes and 5400 exomes Using ANNOVAR summary routine I have got a exome_summary.csv file that contains different columns with 1000 genomes and 5400 exomes frequencies with some empty cells. I asume that empty cells correspond to SNPs that were not found in any of these databases. However there are cases where you get frequencies either in 1000 genomes or 5400 exomes but not in the other database. I would understand it if the alllele frequency in the alternate database would be very low, but there are cases where this frequency can be up to 80%. Does anyone know what is the reason for these big discrepancies between these databases?

10-31-2013, 11:01 AM	#2
etal Member Location: California Join Date: Oct 2013 Posts: 23	The 1000G and 5400ESP projects used different pipelines for sequencing, read mapping and and variant calling, so at least some of the major discrepancies between the listed allele frequencies are due to different quirks in their respective pipelines. For example, each pipeline might consistently map ambiguous reads to a different site in a repetitive or low-complexity region, giving the variant calls in that region a high frequency in one database but not the other.