Changes between Version 6 and Version 7 of SOPs/atac_Seq

Timestamp:

08/11/20 11:17:09 (4 years ago)

Author:

thiruvil

Comment:

--

SOPs/atac_Seq

@@ -85 +85 @@
 //Based on presentation by H.Liu at BioC 2020 presentation //
   *  post-alignment filtering: remove mito/ChrM reads, remove duplciates, remove mapping artifacts: < 38bp and fragments > 2kb, and discordant reads
-    * The NRF should be > 38bp to < 147 bases (one nucleosome); mono nucleosome fragments are in the range of 147-200 bp
-    * don't recommend to remove di-, tri nucleosome fragments for downstream analysis unless you have a great number of reads. Otherwise the you might miss many open chromatin regions during peak calling.
+    * The nucleosome free region (NFR) should be > 38bp to < 147 bases (one nucleosome); mono nucleosome fragments are in the range of 147-200 bp
+    * unless you have high coverage, e.g. ~50M reads or more, use all reads for downstream analysis and not remove mono-, di-, tri-, etc. nucleosome fragments or reads.  Otherwise, you may miss many open chromatin regions during peak calling.
   * check fragment size distribution (from ATACSeqQC)
     * see [[https://www.nature.com/articles/nmeth.2688/figures/2 | Fig 2]] from Buenrostro, J.D., et al. for expected distribution of fragment size
-  * Tn5 produces 5’ overhangs of 9 bases long: pos. strand + 4 and neg strand -5 (shiftGAlignmentsList and shiftReads function) splitGcAlignmentsByCut: creates different bins of reads: nucleosome free regions (NFR), mono, di, etc. Shifted reads that do not fit into any of the bins should be discarded.
+  * Tn5 produces 5’ overhangs of 9 bases long: pos. strand + 4 and neg strand -5 (shiftGAlignmentsList and shiftReads function) splitGcAlignmentsByCut: creates different bins of reads: NFR, mono, di, etc. Shifted reads that do not fit into any of the bins should be discarded.
   *  use housekeeping genes to check QC: signal enrichment is expected in the regulatory regions of housekeeping genes in good ATACSeq eperiments. Use IGVSnapshot function with geneNames param. splitGAlignmentsByCut: creates different bins of reads: NFR, mono, di, etc. Shifted reads that do not fit into any of the bins should be discarded.