Egrl is a fresh implementation of e-genotype, an allele screening method for Next-Gen sequencing data.
Here's some of the new features:
-
probes and reads can be feed in the stdin. Yes, who doesn't love Unix? This also allows the processing of data from bam/sam.
-
Fasta(q) allowed.
-
Processes compressed files (gzip).
-
MultiThreading support.
-
Variable flanking sequence length.
-
New execution mode that reports hits as they are seen.
$ git clone git@github.com:drio/egrl.git
$ cd egrl && sh ./autogen.sh && ./configure && make && make install
$ egrl
Now let's say we have some illumina reads:
$ head reads.fata
>1
GGTTTACACTCTAGGTTACCGTGGGGGAAAGACTG
>2
GTTATCCAAGCTTCAGCAAAATATATCAGCATCAT
>3
GATTAATAACATGTTGATCTTTTTCTTCTTTCTGT
>4
GGTTTTCTAGAGGCAGAAGGCCAAGTTCATTCTCT
>5
GAGGGTTGATTCCAGAGAAGAACAAGTAGCCTCTC
And also we have some probes:
$ head probes.txt
1 100006955 rs4908018 TTTGTCTAAAACAAC CTTTCACTAGGCTCA C A
1 100007331 rs2392072 TTATCATTCCCTTCC GATCACCTCTACCAG A G
1 100008014 rs7541580 CTTCAGCTGAGAATG TAAGGACCTGTGTGG T C
1 100008945 rs903127 GGCTGACAAAGACAC GAGCTAGCAAATGAG G A
1 100016483 rs4907889 TGGGCTGGCCATAAA CAATCAAAACCTCCT C T
1 100018883 rs12728909 GGGACTGAGGTAAGC GTAGGGAAAGGTGGA C A
1 100024118 rs17121193 AAGTACACAATGTCC AAGGCCTTTTTCATT G A
1 100024805 rs10875258 AGTGTAGTCCACCAA CAGTAATATCAGCAT T C
1 100040066 rs12023333 GTTGCTTAACAATCC TTCTTCAGTTCATCT T C
1 100040996 rs6688707 TTCAGGCCTTCAGTG ATTTCCATGAGACCC G T
Probes is a tab delimited file that contains information of a particular locus in the genome. The fields are: chromosome, coordinate, id, 5' flanking sequence, 3' flanking sequence, expected reference allele and expected variant allele.
Now we can use egrl to count how many times any of these probes
are being hit by our reads. I am assuming egrl is in your $PATH
$ egrl count -p probes.affy.txt -r reads.fasta -t 8 > counts.csv
egrl dumps stuff to the stderr and the "counts" to the standard output:
$ head counts.csv
3,187080482,rs6809601,T,C,0,0,0,0,0,0,1,0,0,0,0,1,0,0,0
9,83526867,rs1471723,G,C,0,0,0,0,0,0,1,0,0,0,0,1,0,0,0
1,193412546,rs599140,T,C,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0
22,23680087,rs4140486,A,G,1,0,0,0,0,1,0,0,0,0,2,0,0,0,0
1,149638762,rs7172,G,A,0,1,0,0,0,0,0,0,0,0,0,1,0,0,0
4,170686155,rs10010203,A,G,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0
3,187826181,rs13080283,G,A,2,0,0,0,0,0,0,0,0,0,2,0,0,0,0
4,8985664,rs9291627,T,C,0,0,0,0,0,0,1,0,0,0,0,1,0,0,0
13,103914442,rs7322187,G,A,0,0,0,0,0,1,0,0,0,0,1,0,0,0,0
8,4567331,rs17346747,G,A,0,0,0,0,0,0,1,0,0,0,0,1,0,0,0
The output is quite simple.A csv file with the chromosome, coordinate, probe id, reference, variant and then the allele counts from the reads. Only probes being hit are displayed.
The allele counts are separated in three parts: hits to the ref allele, to the variant, to the alternative allele1 and the alternative allele2. The next five columns are the same but from hits to the reverse complement of the probe. The last five columns add the hits from the previous columns.
There is another useful command that enumerates all the hits to any of the probes. As soon as we find a hit, this is dumped to the stdin. Each hits consists on the probe id, the allele seen and the read id.
egrl hits uses much less memory since we are not storing in memory
the information related to a hit. Notice also we are reporting the read
that hit the probe. You may find that useful.
Here you have a running example against an illumina dataset.
$ egrl hits -p tests/probes.small.txt -r tests/reads.small.fq
>> Loading probes
[timer - Probe loading (SS)] wall clock: 0.23s CPU: 0.23s
>> # of probes (RC included): 40000
rs8067076,A,DCT4KXP1:297:C196LACXX:1:1101:3921:2216
rs2073067,G,DCT4KXP1:297:C196LACXX:1:1101:7679:2092
rs1065767,G,DCT4KXP1:297:C196LACXX:1:1101:8842:2157
rs10509305,A,DCT4KXP1:297:C196LACXX:1:1101:16274:2067
rs4844831,G,DCT4KXP1:297:C196LACXX:1:1101:7721:2569
rs7093194,G,DCT4KXP1:297:C196LACXX:1:1101:7698:2966
rs2272379,A,DCT4KXP1:297:C196LACXX:1:1101:9899:3023
[timer - count_main] wall clock: 0.46s CPU: 0.46s