1
$\begingroup$

Using the SRAdb R package, I am looking to query individuals who possess both WGS and RNA-Seq data on the SRA. So far I believe I have been able to query specific samples with both WGS and RNA-Seq experiments, however a specific sample, provided by the sample_accession, is not exactly what I'm looking for.

For example: I'm interested in finding human data in which one human was used to derive WGS data, and several tissues from that same human were used to derive RNA-Seq data.

Is there any easy way to accomplish this using SQL and/or R?

So far, this is my approach to identify samples that have both WGS and RNA-Seq data, using the local database, SRAmetadb.sqlite:

con <- dbConnect(SQLite(),'SRAmetadb.sqlite')
query <- dbGetQuery(con,
                    paste(
                          "SELECT
                            sample.sample_accession,
                            sample.scientific_name,
                            experiment.experiment_accession,
                            experiment.library_strategy
                          FROM sample
                          JOIN experiment ON
                            sample.sample_accession = experiment.sample_accession
                          WHERE experiment.library_strategy in ('WGS','RNA-Seq')
                          GROUP by sample.sample_accession
                          HAVING COUNT(DISTINCT experiment.library_strategy) = 2"
                        )
                    )

With this approach, I can see query contains 181 human samples that have at least one WGS experiment and one RNA-Seq experiment. That's great, but I feel I must be missing out on data that came from different samples but the same individuals.

$\endgroup$
  • $\begingroup$ Why do you feel so? $\endgroup$ – Maxim Kuleshov Jul 13 '16 at 22:20
  • $\begingroup$ A liver sample and a brain sample from the same individual would have different sample_accession values, correct? $\endgroup$ – Scott Funkhouser Jul 13 '16 at 22:24
  • $\begingroup$ And it's remarkable to me that there are only 4 Mus musculus samples that meet this criteria, using the method above $\endgroup$ – Scott Funkhouser Jul 14 '16 at 20:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.