1. Virus discovery-454 sequencing
Michel de Vries
Laboratory of
Experimental Virology

2. Introduction
In 5-40 % of hospitalized patients with suspected respiratory viral infection no agent is identified.
Possible problem:
New or unusual subtype.
Virus Discovery cDNA-AFLP (VIDISCA) was developed in *
VIDISCA can amplify both RNA and DNA viruses without prior knowledge of the target
sequence.
Based on restriction enzyme cleavage sites + ligation of adaptors + PCR.
* van der Hoek et al, Nature medicine 2004

3. VIDISCA Selective PCR (16 primer combin.) Fragment separation
and isolation
Cloning in TA vector
Colony-PCR
Sequencing of colony-PCR products
Selective PCR (16 primer combin.)

4. Problems
VIDISCA amplifies background ribosomal RNA (rRNA) and chromosomal DNA together with viral sequences.
This background amplification interferes with VIDISCA by acting as competitors.
VIDISCA can only amplify high viral load samples and/or low background samples.
Clinical samples such as nose washing are full with background rRNA and chromosomal DNA.

5. Clinical samples
12 clinical samples (nose washings) supplied by clinical virology.
Virus in the samples were identified via multiplex PCR but given double blind.
Samples containing high/medium/low viral load of known viruses.
For 1 out of 12 samples a viral sequences was obtained (HCoV-229E).
In most samples only background rRNA and DNA was identified

6. Solution If a high number of fragments are randomly sequenced, a
minority population can be identified.

7. Next Generation Sequencing
One of NGS is the 454 sequencer of Roche, which is present in the AMC.
Per run a maximum of 1.5 E6 beads can be used resulting in about quality sequences.
Multiplex identifiers (MID) are 10 nucleotides long barcodes that are recognized by our software.
The MID can be incorporated into samples allowing multiple samples to be pooled.
Primer A MID DNA fragment Primer B

8. VIDISCA- 454 454 sequencing Selective PCR (16 primer combin.)
Fragment separation
and isolation
Cloning in TA vector
Colony-PCR
Sequencing of colony-PCR products
Selective PCR (16 primer combin.)
VIDISCA-
454
454
sequencing

9. VIDISCA-454 Can we sequence viruses with VIDISCA-454 ?
Can we use 12 MIDs in one run ?

10. Test with human coxsackievirus B4
12 times coxsackievirus B4 supernatant (2.0 E8 copies/ml) as input.
Each with a specific MID-primer A anchor.
Sequences were separated per MID, aligned and compared to GenBank database.

11. Result MID Nr. of sequences Viral sequences % of total 1 2805 2477
88.3 2
3890
3282
84.4 3
2369
2007
84.7 4
2782
2534
91.1 5
3056
2540
83.1 6
2271
2000
88.1 7
1429
1354
94.8 8
3830
2961
77.3 9
3018
2821
93.5 10
3040
2840
93.4 11
5216
4991
95.7 12
2891
2740
Total
36597
32547
88.9

12. Result 2 Standard VIDISCA VIDISCA-454 Number of fragments. 11 35
Total number of viral
genome nucleotides
Sequenced.
2646
4365
% of viral genome. 36 59

13. Conclusion VIDSICA-454 works!!
12 MIDs can be used, so 12 samples can be pooled in a sequence run.

14. Are we ready for clinical samples??
Conclusion
VIDSICA-454 works!!
12 MIDs can be used, so 12 samples can be pooled in a sequence run
Are we ready for clinical samples??

15. Result 3 Again the12 clinical samples tested with VIDSICA-454

16. In 6 out of 12 samples a virus could be identified.
Result 4
In 6 out of 12 samples a virus could be identified.
VIDISCA
VIDISCA-454
Samples tested 12
Nr. of virus
indentified 1 6
Viruses identified
HCoV 229E
HCoV OC43
RSV (2X)
Rhinovirus
hMPV

17. Sample ID
VIDISCA
VIDISCA-454
(% positive)
Input
Negative
Negative (<0.014 %)
RSV
HCoV 229E
(1.1 %)
hMPV
(0.006 %)
HCoV OC43
(0.006%)
(< 0.007)
HRV
(0.33 %)
(<0.04 %)
(<0.05%)
HRV-C
(0.017 %)
(<0.012%)
(<0.021 %)
(0.166 %)

18. Conclusion We are ready for clinical samples!!
With VIDISCA out of 12 clinical samples could be identified compared to 1 out of 12 with normal VIDISCA.
An increase of sequence data by 454 sequencing results in a high chance of viral detection in clinical samples.
We are ready for clinical samples!!

19. Advantage VIDISCA-454 More sensitive. Cheaper per sample.
More samples per time unit.
More sequence information per sample.
Less interference by background amplification.
More sensitive.

20. Problems
One of the critical steps is measuring/calculating the amount of DNA.
DNA is measured in ng/µl via fluormeter and DNA size is estimated via agarose gel.
Analyses takes 1 month.
Codoncode (aligning program) can not handle > 2000 sequences. Therefore you have to align in batches.
For analyses via Blast sequences have to be exported in batches of 100 and then submitted.

21. Acknowledgements Laboratory of experimental virology
Nuno Faria
Marta Canuti
Martin Deijs
Maarten F. Jebbink
Lia van der Hoek
Department of Neurogenetics
Marja jakobs
Frank Baas
Deparment of clinical virology
Richard Molenkamp
Klinische Epidemiologie, Biostatistiek en Bioinformatica
Barbera van Schaik
Angela Luijf