APT rma-sketch gene has same values as exon - Affymetrix Human Exon 1.0 ST Array

Question

[bjwiley23]

Hi, I have below code to summarize the exon and gene intensities. I am wondering why for 1 gene it has the same values (logged) as one of the probes so that when doing Gene-level Normalized intensity subtracting (since its logged) I'll get 0 for each sample. How is this the case that gene level intensity returned from APT is same as an exon?

#gene estimates (exon array)
apt-probeset-summarize \
-p /software/apt/lib/HuEx-1_0-st-v2.r2.pgf \
-c /software/apt/lib/HuEx-1_0-st-v2.r2.clf \
-b /software/apt/lib/HuEx-1_0-st-v2.r2.antigenomic.bgp \
--qc-probesets /software/apt/lib/HuEx-1_0-st-v2.r2.qcc \
-m /software/apt/lib/HuEx-1_0-st-v2.r2.dt1.hg18.core.mps \
-a rma-sketch \
--kill-list /software/apt/lib/Affymetrix_HuEx_1.0ST_Affected_probes_probesets.txt \
-o output-dir/gene \
/cels/exon/keep/*.CEL

#exon estimates (exon array)
apt-probeset-summarize \
-p /software/apt/lib/HuEx-1_0-st-v2.r2.pgf \
-c /software/apt/lib/HuEx-1_0-st-v2.r2.clf \
-b /software/apt/lib/HuEx-1_0-st-v2.r2.antigenomic.bgp \
--qc-probesets /software/apt/lib/HuEx-1_0-st-v2.r2.qcc \
-s /software/apt/lib/HuEx-1_0-st-v2.r2.dt1.hg18.core.ps \
-a rma-sketch \
--kill-list /software/apt/lib/Affymetrix_HuEx_1.0ST_Affected_probes_probesets.txt \
-o output-dir/exon \
/cels/exon/keep/*.CEL

An example is transcript_cluster_id 2315918

#gene
grep 2315918 rma-sketch.summary.txt
2315918 6.42343 6.51861 6.41761 6.48735 5.99480 6.52973 6.05458 6.88223 7.26753 6.37890

And one of the exons 2315941

#exon
grep 2315941 rma-sketch.summary.txt
2315941 6.42343 6.51861 6.41761 6.48735 5.99480 6.52973 6.05458 6.88223 7.26753 6.37890

This happened for more than 1 study (above is GSE18300) as well for this exact exon and transcript cluster so is it specific to a particular gene? This is for NM_001039211 // ATAD3C, exon id 439.

Answer 1

[Kevin Blighe]

I am not sure how APT summarises the probe-level intensities but, in this case, 2315941 is a probeset_id that is related to the 2315918 transcript_cluster_id.

This transcript cluster ID has many other probe-set IDs related to it, including:

"probeset_id","seqname","strand","start","stop","probe_count","transcript_cluster_id","exon_id"
"2315920","chr1","+","1385169","1386029","4","2315918","424"
"2315921","chr1","+","1386073","1386138","4","2315918","425"
"2315922","chr1","+","1387426","1387450","1","2315918","426"
"2315923","chr1","+","1387750","1387814","4","2315918","427"
"2315926","chr1","+","1389542","1389630","4","2315918","428"
"2315927","chr1","+","1389749","1389852","4","2315918","429"
"2315928","chr1","+","1390851","1390898","4","2315918","430"
"2315929","chr1","+","1391192","1391220","2","2315918","431"
"2315930","chr1","+","1391606","1391725","4","2315918","432"
"2315933","chr1","+","1392523","1392558","4","2315918","435"
"2315934","chr1","+","1394542","1394601","4","2315918","436"
"2315935","chr1","+","1395784","1395808","1","2315918","437"
"2315936","chr1","+","1397917","1397968","4","2315918","438"
"2315937","chr1","+","1398015","1398046","4","2315918","438"
"2315938","chr1","+","1398052","1398077","2","2315918","438"
"2315939","chr1","+","1403811","1403838","4","2315918","439"
"2315940","chr1","+","1403973","1404058","4","2315918","439"
"2315941","chr1","+","1404605","1404917","4","2315918","439" <<-- HERE
"2315919","chr1","+","1381389","1381416","4","2315918","423"
"2315924","chr1","+","1387817","1387841","1","2315918","427"
"2315925","chr1","+","1387905","1387957","4","2315918","427"
"2315931","chr1","+","1391886","1391963","3","2315918","433"
"2315932","chr1","+","1392270","1392367","4","2315918","434"

I have put a 'HERE' where your probe-set of interest matches.

Depending on the level of sumarisation that you're doing, you should either get a single value for each probe-set ID, or else a single [summarised] value for the transcript cluster ID.

If you want, I can share R code with you so that you can have more control over this analysis. I'm really not sure how APT processes these.

Kevin