Application for aberration calling in aCGH data

Overview Algorithm Download  Examples References


The Stepgram application is used to detect aberrations in DNA copy number data from aCGH experiments. The application is based on the interval score described in the paper "Efficient Calculation of Interval Scores for DNA Copy Number Data Analysis" (Proceedings of RECOMB 05, and JCB 06),

Features of the Stepgram application:

  • Rapid identification of aberrations in multiple chromsomes/samples (<1 sec for a 10K probe data vector)

  • Identification of aberrations (gains & losses) rather than mere segmentation of the data vector.

  • Aberration calls are made based on deviation of the signal from the zero baseline, not by breakpoint detection.

  • Detection of aberrations consisting down to a single probe.

  • Results are exported in two formats: (i) a list of aberrant intervals, (ii) aberration level per probe for simple plotting and processing in external tools.

Breast cancer cell line sample MDA-MB-453, chromosome 17 (7,723 probes): Smoothed raw data and aberration calls (data from Barrett et al. PNAS 04)


Aberration calling in Stepgram is based on the Interval Score defined in (Lipson et al, RECOMB 05). This score describes, for each genomic interval I, the deviation of the sum of signals in the interval from the zero baseline:

where xi is the i'th elements in the data vector, and is the noise level of the data, measured as the derivative log-ratio spread (DLRS) of the signal over the entire genome. The interval I that maximizes the score S(I) is determined using the efficient LookAhead algorithm. If this score passes a given threshold, then I is marked as aberrant, and the procedure is repeated recursively to the left and right of I, and also within I - to search for nested aberrations. The recursion is stopped when no interval with a score surpassing the given threshold is found.


Stepgram is available as a Win32 executable (version 2.12, tested on Windows XP):

Stepgram application help page


Stepgram output is easily imported to and visualized in external applications, such as Excel, Origin etc:

Breast cancer cell line samples BT474 (top) and SKBR3 (bottom): Raw data (from Pollack et al. PNAS 02) and Stepgram aberration calls. Stepgram applied with parameters threshold=8, minDiff=0.1 (visualization in MS Excel).


If you choose to use this application in your research, please cite:

  • Doron Lipson, Yonatan Aumann, Amir Ben-Dor, Nathan Linial and Zohar Yakhini, "Efficient Calculation of Interval Scores for DNA Copy Number Data Analysis". Journal of Computational Biology, 13(2):215-228, 2006. (Paper)
    (Also in Proceedings of RECOMB '05, LNCS 3500, p. 83, Springer-Verlag, 2005)

Contact Info

For additional Information please contact Doron Lipson, CS Department, Technion, Israel.

hit counter