Revision as of 16:55, 11 December 2013

Information on this page is for version 0.569 or higher. Sorry for confusion, hopefully program and wiki will be updated before weekend.

We allow for filtering at many different levels.

Read level, MapQ, unique mapped reads etc
Base level, qscore
Sequencing depth
Regions (using BAM indexing (active lookup))
Single sites (passive lookup, also allows for forcing major and minor) -sites
Filtering based on downstream analysis. minimum MAF, LRT for SNP calling etc.
Trimming out the ends of the reads
etc

It follows that some filters will select a subset of data, and some of the filters will discard certain sites. If multiple filters has been chosen, the analysis will be limited to the chain of filters. Eg setting a qscore threshold and an effective sample size filter along with a MAF filter will first. Remove the data with low qscores, then we found out the number of samples with data, and remove those below the threshold. Then we calculate the MAF and remove those sites with af MAF below the threshold.

This page will describe some of these filters, and it follows that only some are available in the case of BAM input.

Read level filters

We allow for filtering and manipulation a the read level using the following arguments.

-only_proper_pairs [int]=0

Include only proper pairs (pairs of read with both mates mapped correctly). 1: include only proper (default), 0: use all reads. If your data is not paired end you have to choose 1

-uniqueOnly [int]=0

remove reads that have multiple best hits.. 0 no (default), 1 remove

-remove_bads [int]=1

Same as the samtools flags -x which removes read with a flag above 255 (not primary, failure and duplicate reads)

-minQ [int]=0

minimum base quality

-minMapQ [int]=0

minimum mapQ quality. Internally this is handled by setting the bases with a qscore below the threshold to 'N'. -baq [int] =0 perform baq computation, remember to cite the baq paper for this.

Selected Regions

BAM files allows for indexing which makes random retrieval of regions fast and easy. This section describes region lookup as we have implemented it in angsd.

-r [region]

Specify a region with in a chromosome using the syntax [chr]:[start-stop]. examples

chr1:1-10000             \\ first 10000 based for chr1
chr2:50000-               \\chr2 but exclude the first 50000 bases
chr11:1-                  \\all of chr11
chr7:123456              //position 123456 of chr7

-rf [region file]

specify multiple regions in a file.

The format for the regions supplied to the -rf file is the same the -r command line arguments.

Selected Sites

If you are interested in running your analysis at individual sites that are distributed throughout the entire genome, it might be faster to simply to loop over the entire data, but only analyse the data at specific positions. This can be done by supplying the -sites argument. With this approach we also allows for the forcing of major/minor alleles.

Allele frequencies

-minMaf [float]: only work with sites with a maf above 'float'

Of cause requires -doMaf.

Polymorphic sites

-minLRT [float]: only work with sits with an LRT>float

Number of non missing individuals

-minInd [int]: only work with sites with information from atleast int individiduals

Extra

-setMinDepth

Discard site if sequencing depth is below threshold

-setMaxDepth

Discard site if sequencing depth is above threshold - These filter is implemented in -doCounts.

-geno_minDeph

Only call genotypes if per sample genotypes are above this threshold

This requires -doCounts and -doGeno

Examples

First we do a run with no filters

./angsd  -doMaf 2 -doMajorMinor 1 -out TSK -bam bam.filelist -GL 1 -r 1:
...
head TSK.mafs 
chromo	position	major	minor	knownEM	nInd
1	13999919	A	C	0.000008	1
1	13999920	G	A	0.000008	1
1	13999921	G	A	0.000008	1
1	13999922	C	A	0.000008	1
1	13999923	A	C	0.000008	1
1	13999924	G	A	0.000008	1
1	13999925	G	A	0.000008	1
1	13999926	A	C	0.000008	1
1	13999927	G	A	0.000008	1

Now we do a filter with MAF cutoff of 1\%

../angsd0.3/angsd -doMaf 2 -doMajorMinor 1 -out TSK -bam bam.filelist -GL 1 -r 1: -minMaf 0.01
head TSK.mafs 
chromo	position	major	minor	knownEM	nInd
1	13999950	T	G	0.495291	2
1	14000019	G	T	0.047247	9
1	14000056	C	T	0.055851	10
1	14000127	G	T	0.060760	10
1	14000170	C	T	0.052388	9
1	14000176	G	A	0.047928	10
1	14000202	G	A	0.279722	9
1	14000262	C	T	0.058555	9
1	14000322	A	G	0.040471	8

Similar if we only want sites with information for atleast 5 samples

../angsd0.3/angsd -doMaf 2 -doMajorMinor 1 -out TSK -bam bam.filelist -GL 1 -r 1: -minKeepInd 5
head TSK.mafs 
chromo	position	major	minor	knownEM	nInd
1	13999971	T	A	0.000007	6
1	13999972	G	A	0.000007	6
1	13999973	C	A	0.000005	5
1	13999974	G	A	0.000006	6
1	13999975	C	A	0.000002	5
1	13999976	C	A	0.000004	7
1	13999977	A	C	0.000005	8
1	13999978	C	A	0.000005	8
1	13999979	T	A	0.000005	8

If we are interested in all sites with a p-value of 10^(-6) of being variable

../angsd0.3/angsd -doMaf 2 -doMajorMinor 1 -out TSK -bam bam.filelist -GL 1 -r 1: -minLRT 24 -doSNP 1 
head TSK.mafs 
chromo	position	major	minor	knownEM	pK-EM	nInd
1	14000202	G	A	0.279722	42.623150	9
1	14000873	G	A	0.212120	79.118476	10
1	14001018	T	C	0.333736	89.040311	8
1	14001867	A	G	0.200232	47.195423	10
1	14002422	A	T	0.167692	43.196259	9
1	14003581	C	T	0.207404	58.593208	9
1	14004623	T	C	0.219838	102.856433	10
1	14007493	A	G	0.453217	28.398647	9
1	14007558	C	T	0.395670	80.236777	7

@@ Line 52: / Line 52: @@
 =Allele frequencies=
 ; -minMaf [float]: only work with sites with a maf above 'float'
+Of cause requires [[Allele Frequency estimation | -doMaf]].
 =Polymorphic sites=

Filters: Difference between revisions