diff --git a/python/genvarloader/_dataset/_msa.py b/python/genvarloader/_dataset/_msa.py
new file mode 100644
index 0000000..8e7728b
--- /dev/null
+++ b/python/genvarloader/_dataset/_msa.py
@@ -0,0 +1,256 @@
+import time
+import numpy as np
+import numba as nb
+
+
+@nb.njit(cache=True)
+def stack_ploidy(arr):
+    """Optimized ploidy stacking using numba.
+
+    This function takes a 3D array of shape (n_samples, 2, seq_len) containing diploid sequences
+    and stacks them into a 2D array of shape (n_samples*2, seq_len) where each sample's two
+    haplotypes are placed in consecutive rows.
+
+    Args:
+        arr: 3D numpy array of shape (n_samples, 2, seq_len) containing diploid sequences
+            where arr[i,0] is the first haplotype and arr[i,1] is the second haplotype
+            for sample i.
+
+    Returns:
+        2D numpy array of shape (n_samples*2, seq_len) where rows 2*i and 2*i+1 contain
+        the two haplotypes for sample i.
+    """
+    n_samples = arr.shape[0]
+    seq_len = arr.shape[2]
+    result = np.empty((n_samples * 2, seq_len), dtype=arr.dtype)
+    for i in range(n_samples):
+        result[i * 2] = arr[i, 0]
+        result[i * 2 + 1] = arr[i, 1]
+    return result
+
+
+@nb.njit(parallel=True, cache=True)
+def _create_msa_fast(ref_coords, seq_arr, min_coord, max_coord):
+    """Numba-accelerated core MSA creation function.
+
+    This internal function creates a gappy Multiple Sequence Alignment (MSA) by mapping
+    sequence positions to their corresponding reference coordinates. It uses Numba's
+    parallel processing capabilities for improved performance.
+
+    Args:
+        ref_coords: 2D array of shape (n_seqs, seq_len) containing reference coordinates
+            for each position in each sequence
+        seq_arr: 2D array of shape (n_seqs, seq_len) containing the sequence data as
+            integer-encoded nucleotides
+        min_coord: Integer indicating the minimum reference coordinate across all sequences
+        max_coord: Integer indicating the maximum reference coordinate across all sequences
+
+    Returns:
+        2D array of shape (n_seqs, alignment_length) containing the gappy MSA where:
+        - Each row represents one sequence
+        - Each column represents one position in the reference
+        - Gaps are represented by the ASCII value of '-'
+        - Nucleotides are represented by their ASCII values
+    """
+    n_seqs = ref_coords.shape[0]
+    alignment_length = max_coord - min_coord + 1
+    msa = np.full((n_seqs, alignment_length), ord("-"), dtype=np.int8)
+
+    # Pre-compute coordinate offsets to avoid repeated subtraction in the loop
+    coord_offsets = ref_coords - min_coord
+
+    for i in nb.prange(n_seqs):
+        coords = coord_offsets[i]
+        seq = seq_arr[i]
+        seq_len = len(coords)
+        for j in range(seq_len):
+            msa[i, coords[j]] = seq[j]
+
+    return msa
+
+
+def create_msa(ref_coords, seq_arr):
+    """
+    Create a gappy Multiple Sequence Alignment from a ragged 2D array using reference coordinates.
+
+    This will take a 2D array of the reference coordinates:
+        ref_coords = [
+        [100, 101, 103],  # First sequence has bases at positions 100, 101, 103
+        [100, 102, 103]   # Second sequence has bases at positions 100, 102, 103
+        ]
+    And a 2D array of the sequences (left-aligned):
+        seq_arr = [
+            ['A', 'T', 'G'],  # First sequence
+            ['A', 'C', 'G']   # Second sequence
+        ]
+    And return a gappy Multiple Sequence Alignment that is aligned at all positions.
+        A T - G  # First sequence
+        A - C G  # Second sequence
+    Args:
+        ref_coords: 2D array of reference coordinates (sequence x position) indicating where each position maps in the reference genome
+        seq_arr: 2D array of sequences (sequence x position) with ragged right end
+
+    Returns:
+        2D array containing the gappy MSA
+    """
+    start_time = time.time()
+
+    assert ref_coords.shape == seq_arr.shape, (
+        f"Shapes do not match: ref_coords.shape ({ref_coords.shape}) != seq_arr.shape ({seq_arr.shape})"
+    )
+
+    # Convert input arrays to numpy arrays if they aren't already
+    ref_coords = np.asarray(ref_coords)
+    seq_arr = np.asarray(seq_arr)
+
+    # Find the min and max reference coordinates across all sequences
+    min_coord = np.min(ref_coords)
+    max_coord = np.max(ref_coords)
+
+    # Optimize byte conversion using vectorized operations
+    if seq_arr.dtype == np.dtype("|S1"):
+        seq_arr_int = np.frombuffer(seq_arr.tobytes(), dtype=np.int8).reshape(
+            seq_arr.shape
+        )
+    else:
+        seq_arr_int = np.array(
+            [[ord(c) for c in row] for row in seq_arr], dtype=np.int8
+        )
+
+    # Create MSA using numba-accelerated function
+    msa = _create_msa_fast(ref_coords, seq_arr_int, min_coord, max_coord)
+
+    # Convert back to bytes array using view
+    msa = msa.view("|S1")
+
+    end_time = time.time()
+    print(
+        f"MSA for {seq_arr.shape[0]} sequences done {end_time - start_time:.2f} seconds"
+    )
+
+    return msa
+
+
+def preview_msa(msa, max_len=None):
+    """
+    Preview the MSA by printing the sequences with differences.
+
+    Args:
+        msa: 2D array containing the gappy MSA
+        max_len: Maximum number of columns to print
+
+    Returns:
+        None
+    """
+    # Convert byte arrays to strings and find columns with differences using numpy vectorization
+    msa_str = np.array(
+        [[x.decode("utf-8") for x in row] for row in msa], dtype="U1"
+    )  # Convert bytes to unicode strings row by row
+
+    # Find columns with differences using numpy operations
+    col_unique = np.apply_along_axis(lambda x: len(np.unique(x)), 0, msa_str)
+    diff_cols = np.where(col_unique > 1)[0]
+
+    # Limit the number of columns if max_len is specified
+    if max_len is not None and len(diff_cols) > max_len:
+        diff_cols = diff_cols[:max_len]
+
+    # Print sequences showing only columns with differences using numpy indexing
+    for seq in msa_str:
+        print("".join(seq[diff_cols]))
+
+
+@nb.njit(nogil=True, cache=True)
+def _get_consensus(msa_uint8):
+    """
+    Get the consensus sequence from a multiple sequence alignment (MSA) represented as uint8 array.
+
+    For each column in the MSA, this function finds the most frequently occurring value.
+    The function is accelerated using Numba for better performance.
+
+    Args:
+        msa_uint8 (np.ndarray): 2D array of shape (n_sequences, n_columns) containing the MSA
+                               encoded as uint8 values.
+
+    Returns:
+        np.ndarray: 1D array of length n_columns containing the consensus sequence
+                    encoded as uint8 values.
+    """
+    n_cols = msa_uint8.shape[1]
+    consensus = np.zeros(n_cols, dtype=np.uint8)
+
+    for col in range(n_cols):
+        # Count occurrences of each value
+        counts = np.zeros(256, dtype=np.int32)
+        for row in range(msa_uint8.shape[0]):
+            val = msa_uint8[row, col]
+            counts[val] += 1
+
+        # Find most common value
+        max_count = -1
+        max_val = 0
+        for val in range(256):
+            if counts[val] > max_count:
+                max_count = counts[val]
+                max_val = val
+
+        consensus[col] = max_val
+
+    return consensus
+
+
+def create_consensus_seq(msa, as_str=False):
+    """
+    Create a consensus sequence from a gappy MSA.
+
+    This function takes a multiple sequence alignment (MSA) and generates a consensus sequence
+    by finding the most frequently occurring character at each position. The function uses
+    Numba-accelerated operations for improved performance.
+
+    Args:
+        msa (np.ndarray): 2D array containing the gappy MSA created with create_msa.
+                         Should be a structured array with string dtype.
+        as_str (bool, optional): Whether to return the consensus sequence as a string.
+                               If False, returns a structured array with string dtype.
+                               Defaults to False.
+
+    Returns:
+        np.ndarray or str: The consensus sequence. If as_str is True, returns a string.
+                          Otherwise returns a structured array with string dtype ('|S1').
+    """
+    start_time = time.time()
+    # Convert to uint8 and reshape
+    msa_uint8 = msa.view(np.uint8).reshape(msa.shape[0], -1)
+
+    # Get consensus using numba function
+    consensus_seq = _get_consensus(msa_uint8)
+
+    end_time = time.time()
+    print(
+        f"Consensus sequence of {msa.shape[0]} sequences created in {end_time - start_time:.2f} seconds"
+    )
+
+    if as_str is True:
+        consensus_seq = bytearray_to_string(consensus_seq)
+    else:
+        consensus_seq = consensus_seq.view("|S1")
+    return consensus_seq
+
+
+def bytearray_to_string(byte_arr):
+    """
+    Convert a bytearray to a string.
+
+    Args:
+        byte_arr: bytearray, the bytearray to convert.
+
+    Returns:
+        str, the string of the bytearray.
+
+    Example:
+        >>> bytearray_to_string(np.array([b'A', b'C', b'G', b'T']))
+        'ACGT'
+    """
+    if isinstance(byte_arr, str):
+        return byte_arr
+    return byte_arr.tobytes().decode()