Thread by @BillyBostickson

The problem is more re bioinformatics software used by Chinese Team to assess quality of SC2 data prior to submission

Failure to do so may have led to the inadvertent sharing of poor quality, contaminated or otherwise misleading SC2 data

@stevenemassey @WashburneAlex @Daoyu15

https://twitter.com/BallouxFrancois/status/1636614715236315136

Did Gao et al. use recommended bioinformatic software to prepare & submit SC2 Sample Data, or even to assess data quality prior to submission to GISAID?

I think not!

But deluded & agenda driven western virologists lap up their raccoon dog's vomit & declare the mystery is solved

GISAID only accepts consensus sequences of assembled genomes, not ‘raw’ sequence data

"Although data in GISAID is considered open, access is restricted to individuals with verified accounts, access to the data is thus more likely to become compromised"

covid19dataportal.se/support_servic…

Some problems with Kraken 2 (1)

It breaks the reads into all the possible kmers.

A perfect hit will have all the kmers mapped, if you have one mismatch it will impact all kmers that contain this nucleotide.

Some problems with Kraken 2 (2)

The confidence score (Gao et al say confidence 0.05) is basically the number of matching kmers divided by the total number of kmers. So, Kraken 2 can be used for contigs but other recommended methods which will give more precise results.

Some problems with Kraken 2 (3)

Default Kraken2 downloads mask low complexity genome sequences (it removes these low complexity genomes) before making the database.

So depending on how many repetitive sequences are in a genome, a certain number of kmers will be unclassified.

Some problems with Kraken 2 (4)

you are going to get a lot of false positives with those parameters!

Yes you get more classified but you cant trust the result.

Google "kraken2 --confidence" to see the discussion and judge for yourself.

github.com/DerrickWood/kr…

Some problems with Kraken 2 (5)

The serious issues in microbiome data analysis based on sequence abundances, typically produced by DNA-to-DNA methods like Kraken 2 underestimate (or overestimate) the relative abundances of microbes with smaller (or larger) genome sizes.

Some problems with Kraken 2 (6)

Without careful consideration, these issues will

"fundamentally affect differential abundance analyses & other analytical methods that rely on accurate taxon counts in their input contingency matrix"

ncbi.nlm.nih.gov/pmc/articles/P…

Some problems with Kraken 2 (7)

"While correctly detecting a similar number of human reads, methods based on taxonomic classification, such as Kraken2 often misclassify bacterial reads as human, although the extent of this was species-specific"

microbiologyresearch.org/docserver/full…

Conclusion

I believe it is irresponsible to base significant claims on unverified & deleted data provided by researchers with a well known COI (CCDC, Gao et al.) who may not have used recommended bioinformatics software before uploading their putative data.

Would you agree?

or in other words...Garbage in, Garbage Out

especially in the @TheAtlantic

@katherineykwu @stevenemassey @BallouxFrancois @WashburneAlex @Daoyu15
@MichaelWorobey @flodebarre

Note:

Gao et al declared they use Kraken 2 in their pre-print

assets.researchsquare.com/files/rs-13703…

A nasty example of "access to GISAID data is thus more likely to become compromised"

https://twitter.com/BillyBostickson/status/1636661977388027905

It's a miracle. My #GISAID account, twice disabled without explanation, has been turned back on. Again without explanation @pricklyresearch

https://twitter.com/pricklyresearch/status/1636850610921193473