SLIDE 1 HxV resistance testing in the clinical routine
Alexander Thielen
Arevir 2016
SLIDE 2 in 2015: – >3300 HIV samples (3000 PR/RT, 1800 IN, 750 ENV)
- ~800 for own clinical routine
- ~600 for studies
– >1400 HCV samples (450 NS3, 1000 NS5): – 200 HBV samples additional targets: – HEV, HDV, JCV, RSV, CMV, ...
some numbers
SLIDE 3
– contamination detection – database lookup – covariation analyses – reporting – Apobec filtering post-processing modules
SLIDE 4 – searches for mutation patterns in data from the last 31 days – approach:
- check for all amino acid mutations with ≥10% prevalence in
sample A if they occur with ≥ 2% prevalence in sample B (and vice versa)
- if the number of differences is below a specified cutoff, then
compare the sequences on nucleotide level – provides also information on used barcodes
contamination detection
A B K103N: 90% 5% M184V 80% 0.2% CGATGAACRAGGGA CGATRAACYAGAGA ........X..X..
SLIDE 5
database lookup
SLIDE 6
Covariation analysis
SLIDE 7
reporting
SLIDE 8
– amount of samples with low viral loads increasing – desire to switch under successful therapy – problem: Apobec 3F/3G mutations (G to A) in DNA, what is really there? – approach: naïve Bayes classifiers (related to Reuman et al., 2010) Apobec filter
SLIDE 9
SLIDE 10
– trained on >100mio reads – tested on 523 samples (111 DNA, 412 RNA) – minor problems with subtypes A and C – integrase still in evaluation Apobec filter
hypermutated reads 5% 10% 15% 20% 30% DNA 19.82% 12.61% 11.71% 9.01% 4.50% RNA 15.05% 5.58% 1.46% 0.73% 0.00%
SLIDE 11 – implemented in java – faster, especially for samples with several targets – better procedures for noise reduction – additional information and features – validated on 618 samples from the routine
- comparison of amino acid frequencies (positions with coverage
>1000
deeptypeHIV (version 2.0)
SLIDE 12
Apobec filter (example)
SLIDE 13
Apobec filter (example)
16-22995D (v1) PR-K20I 99.2% PR-G73R 11.1% PR-T74A 10.9% PR-V82I 99.5% PR-L89M 99.5% RT-V90I 99.1% RT-A98S 99.6% RT-M184I 10.2% IN-G140R 18.8% IN-E157K 2.6% IN-G163R 15.9% 16-22995D (v1) (v2) PR-K20I 99.2% PR-G73R 11.1% 11.69% PR-T74A 10.9% 11.73% PR-V82I 99.5% PR-L89M 99.5% RT-V90I 99.1% RT-A98S 99.6% RT-M184I 10.2% 8.86% IN-G140R 18.8% IN-E157K 2.6% IN-G163R 15.9% 16-22995D (v1) (v2) (filtered) PR-K20I 99.2% PR-G73R 11.1% 11.69% 0.01% PR-T74A 10.9% 11.73% 0.65% PR-V82I 99.5% PR-L89M 99.5% RT-V90I 99.1% RT-A98S 99.6% RT-M184I 10.2% 8.86% 8.52% IN-G140R 18.8% IN-E157K 2.6% IN-G163R 15.9%
SLIDE 14
Apobec filter (example)
SLIDE 15
SLIDE 16
Acknowledgments
Kirsten Becker Nina Engel Anna Memmer Bettina Spielberger Martin Däumer Bernhard Thiele
