Why is this insertion detected by NGS a false positive ?

Question

[docfak]

Hello,

We performed whole exome sequencing on a patient with an AmpliSeq Exome kit and the IonTorrent/Life Proton sequencer. Alignment to human genome reference and variant calling are done by IonTorrent plugins. After bioinformatics analysis, we retained one insertion which was infirmed by Sanger sequencing.

I can not determine why this variant is a false positive. Could it be a alignment problem caused by the homopolymers situated some bp after the variant ?

Here is the data :

CHROM X
POS 153667132
REF C
ALT CG
QUAL 98.7837
FILTER PASS
AF 0.84375
AO 23
CG 2.000
CP 0.252
DP 33
DSP 22
FAO 27
FD unknown
FDP 32
FG frameshift
FR .
FRO 5
FSAF 21
FSAR 6
FSRF 5
FSRR 0
FWDB 0.0947456
FXX 0.0302939
GL GDI1
HRUN 0
LEN 1
MILD 25.3658
OALT G
PB 0.5
PBP 1
QD 12.348
RBI 0.232263
REFB -0.107327
REVB 0.212059
RO 8
SAF 17
SAR 6
SRF 8
SRR 0
SSEN 0
SSEP 0
SSSB -0.175701
STB 0.553155
STBP 0.288
TYPE ins
VARB 0.0524829
DT 0
HS 0

And some screens from IGV :

Many thanks in advance for your help.

[Vivek]

From the IGV views, it appears that there's so much activity around there which usually means a poor mapping quality or base qualities along the read. I did check to see if the region was in some kind of a duplication/repeat rich region but that doesn't appear to be the case from the UCSC browser.

[docfak]

Thank you for your quick answer !

[lh3]

IonTorrent is just bad at tiny indels. I bet it is a sequencing problem rather than an informatics problem.