Alu Retrotransposon Matches Seem Artefactual

[DarioS]

I am interested in LINE1, Alu and SVA retrotransposons and I was looking at Alu specifically by INSRMRC=SINE/Alu If I look at the popular software in the retrotransposon research community MELT and look inside of its references folder, I see a FASTA file named ALU.fa which has

> ALU
GGCCGGGCGCGGTGGCTCACGCCTGTAATCCCAGCACTTTGGGAGGCCGAGGCGGGCGGA
TCACGAGGTCAGGAGATCGAGACCATCCTGGCTAACACGGTGAAACCCCGTCTCTACTAA
AAATACAAAAAATTAGCCGGGCGTGGTGGCGGGCGCCTGTAGTCCCAGCTACTCGGGAGG
CTGAGGCAGGAGAATGGCGTGAACCCGGGAGGCGGAGCTTGCAGTGAGCCGAGATCGCGC
CACTGCACTCCAGCCTGGGCGACAGAGCGAGACTCCGTCTC

as the reference sequence of Alu elements.

However, GRIDSS is labelling repeats of As or Ts as Alu, although they don't appear in the reference sequence.

chr2	179636891	gridss42f_35810b	T	TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT.	1033.21	PASS	AS=0;ASC=1X;ASQ=0.00;ASRP=0;ASSR=0;BA=2;BANRP=0;BANRPQ=0.00;BANSR=0;BANSRQ=0.00;BAQ=348.43;BASRP=13;BASSR=4;BEALN=chr1:26810050|-|35M|0;BEID=asm8-264360,asm8-264365;BEIDH=-1,-1;BEIDL=0,0;BMQ=60.00;BMQN=60.00;BMQX=60.00;BQ=1033.21;BSC=4;BSCQ=75.15;BUM=29;BUMQ=609.62;BVF=31;CAS=0;CASQ=0.00;CQ=1033.21;EVENT=gridss42f_35810;IC=0;INSRMP=0.800;INSRMRC=SINE/Alu;INSRMRO=-;INSRMRT=AluSx4;IQ=0.00;LOCAL_LINKED_BY=.;PURPLE_AF=0.253;PURPLE_CN=3.78;PURPLE_CN_CHANGE=0.957;PURPLE_JCN=0.957;RAS=0;RASQ=0.00;REF=123;REFG=GTTTTTTTCTTATTGCATTGG;REFPAIR=147;REMOTE_LINKED_BY=.;RP=0;RPQ=0.00;SB=0.5;SC=1X;SR=0;SRQ=0.00;SVTYPE=BND;TAF=0.170;VF=0	GT:ASQ:ASRP:ASSR:BANRP:BANRPQ:BANSR:BANSRQ:BAQ:BASRP:BASSR:BQ:BSC:BSCQ:BUM:BUMQ:BVF:CASQ:IC:IQ:QUAL:RASQ:REF:REFPAIR:RP:RPQ:SR:SRQ:VF	.:0.00:0:0:0:0.00:0:0.00:0.00:0:0:0.00:0:0.00:0:0.00:0:0.00:0:0.00:0.00:0.00:56:63:0:0.00:0:0.00:0	.:0.00:0:0:0:0.00:0:0.00:348.43:13:4:1033.21:4:75.15:29:609.62:31:0.00:0:0.00:0.00:0.00:67:84:0:0.00:0:0.00:0
chr3	76896610	gridss56b_29622b	A	.AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA	523.93	PASS	AS=0;ASC=1X;ASQ=0.00;ASRP=0;ASSR=0;BA=1;BANRP=0;BANRPQ=0.00;BANSR=0;BANSRQ=0.00;BAQ=0.00;BASRP=0;BASSR=0;BEALN=chr6:96655299|-|37M|0;BEID=asm11-549910;BEIDH=-1;BEIDL=427;BMQ=59.22;BMQN=39.00;BMQX=60.00;BQ=547.09;BSC=2;BSCQ=40.43;BUM=24;BUMQ=506.66;BVF=24;CAS=0;CASQ=0.00;CQ=547.09;EVENT=gridss56b_29622;IC=0;INSRMP=1.00;INSRMRC=SINE/Alu;INSRMRO=+;INSRMRT=AluY;IQ=0.00;LOCAL_LINKED_BY=dsb71;PURPLE_AF=0.280;PURPLE_CN=2.94;PURPLE_CN_CHANGE=0.939;PURPLE_JCN=0.823;RAS=0;RASQ=0.00;REF=121;REFG=CGGTTCAGTTAAAAAAGTTAG;REFPAIR=106;REMOTE_LINKED_BY=.;RP=0;RPQ=0.00;SB=0.0;SC=1X;SR=0;SRQ=0.00;SVTYPE=BND;TAF=0.172;VF=0	GT:ASQ:ASRP:ASSR:BANRP:BANRPQ:BANSR:BANSRQ:BAQ:BASRP:BASSR:BQ:BSC:BSCQ:BUM:BUMQ:BVF:CASQ:IC:IQ:QUAL:RASQ:REF:REFPAIR:RP:RPQ:SR:SRQ:VF	.:0.00:0:0:0:0.00:0:0.00:0.00:0:0:23.16:0:0.00:1:23.16:1:0.00:0:0.00:0.00:0.00:64:52:0:0.00:0:0.00:0	.:0.00:0:0:0:0.00:0:0.00:0.00:0:0:523.93:2:40.43:23:483.49:23:0.00:0:0.00:0.00:0.00:57:54:0:0.00:0:0.00:0
chr6	86583822	gridss114b_35984b	A	.AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA	1400.08	PASS	AS=0;ASC=1X;ASQ=0.00;ASRP=0;ASSR=0;BA=2;BANRP=0;BANRPQ=0.00;BANSR=0;BANSRQ=0.00;BAQ=333.07;BASRP=9;BASSR=8;BEALN=chr19:11998315|-|36M|0;BEID=asm22-865165,asm22-865185;BEIDH=-1,-1;BEIDL=36,282;BMQ=59.59;BMQN=45.00;BMQX=60.00;BQ=1400.08;BSC=9;BSCQ=163.09;BUM=43;BUMQ=903.92;BVF=48;CAS=0;CASQ=0.00;CQ=1380.87;EVENT=gridss114b_35984;IC=0;INSRMP=1.00;INSRMRC=SINE/Alu;INSRMRO=+;INSRMRT=AluSp;IQ=0.00;LOCAL_LINKED_BY=dsb149;PURPLE_AF=0.408;PURPLE_CN=4.70;PURPLE_CN_CHANGE=1.83;PURPLE_JCN=1.92;RAS=0;RASQ=0.00;REF=98;REFG=TGATTAGCTTAATAATTAACC;REFPAIR=119;REMOTE_LINKED_BY=.;RP=0;RPQ=0.00;SB=0.5294118;SC=1X;SR=0;SRQ=0.00;SVTYPE=BND;TAF=0.293;VF=0	GT:ASQ:ASRP:ASSR:BANRP:BANRPQ:BANSR:BANSRQ:BAQ:BASRP:BASSR:BQ:BSC:BSCQ:BUM:BUMQ:BVF:CASQ:IC:IQ:QUAL:RASQ:REF:REFPAIR:RP:RPQ:SR:SRQ:VF	.:0.00:0:0:0:0.00:0:0.00:0.00:0:0:0.00:0:0.00:0:0.00:0:0.00:0:0.00:0.00:0.00:45:56:0:0.00:0:0.00:0	.:0.00:0:0:0:0.00:0:0.00:333.07:9:8:1400.08:9:163.09:43:903.92:48:0.00:0:0.00:0.00:0.00:53:63:0:0.00:0:0.00:0
chr13	96738568	gridss217f_27736b	T	TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT.	118.95	PASS	AS=0;ASC=1X;ASQ=0.00;ASRP=0;ASSR=0;BA=1;BANRP=0;BANRPQ=0.00;BANSR=0;BANSRQ=0.00;BAQ=59.48;BASRP=0;BASSR=3;BEALN=chr7:145453112|+|41M|0;BEID=asm43-255562;BEIDH=-1;BEIDL=0;BMQ=60.00;BMQN=60.00;BMQX=60.00;BQ=118.95;BSC=3;BSCQ=59.48;BUM=0;BUMQ=0.00;BVF=3;CAS=0;CASQ=0.00;CQ=331.31;EVENT=gridss217f_27736;IC=0;INSRMP=1.00;INSRMRC=SINE/Alu;INSRMRO=-;INSRMRT=AluJo;IQ=0.00;LOCAL_LINKED_BY=dsb384;PURPLE_AF=0.028;PURPLE_CN=3.25;PURPLE_CN_CHANGE=0.229;PURPLE_JCN=0.092;RAS=0;RASQ=0.00;REF=148;REFG=AATGCCCTATTATTGTTTAAA;REFPAIR=131;REMOTE_LINKED_BY=.;RP=0;RPQ=0.00;SB=0.6666667;SC=1X;SR=0;SRQ=0.00;SVTYPE=BND;TAF=0.018;VF=0	GT:ASQ:ASRP:ASSR:BANRP:BANRPQ:BANSR:BANSRQ:BAQ:BASRP:BASSR:BQ:BSC:BSCQ:BUM:BUMQ:BVF:CASQ:IC:IQ:QUAL:RASQ:REF:REFPAIR:RP:RPQ:SR:SRQ:VF	.:0.00:0:0:0:0.00:0:0.00:0.00:0:0:0.00:0:0.00:0:0.00:0:0.00:0:0.00:0.00:0.00:59:57:0:0.00:0:0.00:0	.:0.00:0:0:0:0.00:0:0.00:59.48:0:3:118.95:3:59.48:0:0.00:3:0.00:0:0.00:0.00:0.00:89:74:0:0.00:0:0.00:0

[d-cameron]

Thoses are the repeatmasker annotations for the nominal BEALN locations.

However, GRIDSS is labelling repeats of As or Ts as Alu, although they don't appear in the reference sequence.

They do: in hg38 chr1:26810050-26810085 is a poly-A sequence which RepeatMasker annotates as part of a SINE element.

https://genome.ucsc.edu/cgi-bin/hgTracks?db=hg38&lastVirtModeType=default&lastVirtModeExtraState=&virtModeType=default&virtMode=0&nonVirtPosition=&position=chr1%3A26810000%2D26810099&hgsid=880760495_AaWKkAA6yoOeeG3hF6zPnQAvtbco

That said, the above doesn't really help you that much as clearly annotating polyA sequence with SINE isn't a very useful annotation. I'm aware of the problem and am in the processing of adding support for actual RepeatMasker annotations where the inserted sequence is sent off to RepeatMasker. The RepeatMasker output is a pain to parse so I've been putting this bit off as I've had other tasks to do as well.

The overall pipeline is VCF -> (gridss.InsertedSequencesToFasta^) -> RepeatMasker -> (TODO annotate VCF)

The other issue I've had has been the interpretation of multiple hits. What would you expect the correct annotation to be for sequences that have multiple (either overlapping, or no-overlapping) repeatmasker matches?

^ New in the 2.10.0 dev branch.

[DarioS]

... the inserted sequence is sent off to RepeatMasker.

Does that mean GRIDSS will require internet connectivity to make web service queries? That will make it harder for people using HPCs, which often don't allow internet connections on the compute nodes. For example:

$ cat download.pbs
#!/bin/bash
#PBS -q normal
#PBS -l ncpus=1
#PBS -l mem=2gb
#PBS -l walltime=00:03:00

wget ftp://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_35/gencode.v35.annotation.gtf.gz

fails to establish a connection to EBI FTP server

$ cat download.pbs.e10468964
--2020-08-26 09:48:34--  ftp://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_35/gencode.v35.annotation.gtf.gz
           => 'gencode.v35.annotation.gtf.gz'
Resolving ftp.ebi.ac.uk (ftp.ebi.ac.uk)... 193.62.197.74
Connecting to ftp.ebi.ac.uk (ftp.ebi.ac.uk)|193.62.197.74|:21... failed: Connection timed out.
Retrying.

--2020-08-26 09:50:46--  ftp://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_35/gencode.v35.annotation.gtf.gz
  (try: 2) => 'gencode.v35.annotation.gtf.gz'
Connecting to ftp.ebi.ac.uk (ftp.ebi.ac.uk)|193.62.197.74|:21...

Is the functionality available yet in development branch? How would I compile GRIDSS if I wanted to test pre-release on a HPC? Or, is the next release imminent and it would be simplest to just wait for it?

[d-cameron]

Is the functionality available yet in development branch?

Just finished it off today.

How would I compile GRIDSS if I wanted to test pre-release on a HPC

git clone --branch dev http://github.com/PapenfussLab/gridss
cd gridss
 mvn package -DskipTests

Does that mean GRIDSS will require internet connectivity to make web service queries?

No, but it does require a local RepeatMasker installation. I just grabbed the bioconda version.

To test your installation is working:

cd example
./annotate_repeatmasker.sh

[d-cameron]

scripts/gridss_annotate_vcf_repeatmasker.sh will be included in the next release.

[DarioS]

That is convenient but the server which I am using doesn't have RepeatMasker nor conda installed. I'll be busy with dependencies.