skrakau
commented
4 years ago Hi,
in column 5 is the default score at position t, i.e. the log posterior probability ratio of the first and second most likely state, given the observed data D (fragment densities and read start counts): score_(UC:unconditional) = ln(P(S_t = enriched + crosslink | D=d)/P(S_t = 2nd most likely state| D=d))
In column 7 is score_(E:enrichment focused) = ln(P(S_t = enriched + crosslink | D=d)/P(S_t = non-enriched + crosslink| D=d))
score_(CL:crosslink focused) = ln(P(S_t = enriched + crosslink | D=d)/P(S_t = enriched + non-crosslink| D=d))
score_(B:balanced) = ln(P(S_t = enriched | D=d)/P(S_t = non-enriched | D=d)) + ln(P(S_t = crosslink | D=d)/P(S_t = non-crosslink | D=d))
It depends on the binding characteristics of the protein, which score is best. If PureCLIP is run in a mode to correct for biases (input control and CL-motifs) and no prior knowledge is given about the binding characteristics, then the balanced score might be a good choice. Without bias correction I would use the default score.
Hello. Can you please explain what the data in the BED results file mean? In a line like this:
chr20 73636 73637 3 11.54 + [score_CL=25.1365;score_E=11.54;score_B=36.6765;score_UC=11.54]What is the meaning of column 5 (11.54) and the numbers between the square brackets ([score_CL=25.1365;score_E=11.54;score_B=36.6765;score_UC=11.54])?
Thank you.